Fabrics
The following terms are referred to in the document:
-
Greenfield Deployments: Applicable for provisioning new VXLAN EVPN fabrics, and eBGP based Routed fabrics.
-
Brownfield Deployments: Applicable for existing VXLAN EVPN fabrics:
-
Migrate CLI configured VXLAN EVPN fabrics to DCNM using the Easy_Fabric_11_1 fabric template.
-
NFM migration to Cisco DCNM using the Easy_Fabric_11_1 fabric template.
-
For information about upgrades, refer to the Cisco DCNM Installation and Upgrade Guide for LAN Fabric Deployment.
This section contains the following topics:
VXLAN BGP EVPN Fabrics Provisioning
DCNM 11 introduces an enhanced “Easy” fabric workflow for unified underlay and overlay provisioning of VXLAN BGP EVPN configuration on Nexus 9000 and 3000 series of switches. The configuration of the fabric is achieved via a powerful, flexible, and customizable template-based framework. Using minimal user inputs, an entire fabric can be brought up with Cisco recommended best practice configurations, in a short period of time. The set of parameters exposed in the Fabric Settings allow users to tailor the fabric to their preferred underlay provisioning options.
Border devices in a fabric typically provide external connectivity via peering with appropriate edge/core/WAN routers. These edge/core routers may either be managed or monitored by DCNM. These devices are placed in a special fabric called the External Fabric. The same DCNM controller can manage multiple VXLAN BGP EVPN fabrics while also offering easy provisioning and management of Layer-2 and Layer-3 DCI underlay and overlay configuration among these fabrics using a special construct called a Multi-Site Domain (MSD) fabric.
Note that in this document the terms switch and device are used interchangeably.
The DCNM GUI functions for creating and deploying VXLAN BGP EVPN fabrics are as follows:
Control > Fabric Builder menu option (under the Fabrics sub menu).
Create, edit, and delete a fabric:
-
Create new VXLAN, MSD, and external VXLAN fabrics.
-
View the VXLAN and MSD fabric topologies, including connections between fabrics.
-
Update fabric settings.
-
Save and deploy updated changes.
-
Delete a fabric (if devices are removed).
Device discovery and provisioning start-up configurations on new switches:
-
Add switch instances to the fabric.
-
Provision start-up configurations and an IP address to a new switch through POAP configuration.
-
Update switch policies, save, and deploy updated changes.
-
Create intra-fabric and inter-fabric links (also called Inter-Fabric Connections [IFCs]).
Control > Interfaces menu option (under the Fabrics sub menu).
Underlay provisioning:
-
Create, deploy, view, edit and delete a port-channel, vPC switch pair, Straight Through FEX (ST-FEX), Active-Active FEX (AA-FEX), loopback, subinterface, etc.
-
Create breakout and unbreakout ports.
-
Shut down and bring up interfaces.
-
Rediscover ports and view interface configuration history.
Control > Networks and Control > VRFs menu options (under the Fabrics sub menu).
Overlay network provisioning.
-
Create new overlay networks and VRFs (from the range specified in fabric creation).
-
Provision the overlay networks and VRFs on the switches of the fabric.
-
Undeploy the networks and VRFs from the switches.
-
Remove the provisioning from the fabric in DCNM.
Control> Services menu option (under the Fabrics sub menu).
Provisioning of configuration on service leafs to which L4-7 service appliances may be attached. For more information, see L4-L7 Service Basic Workflow.
This chapter mostly covers configuration provisioning for a single VXLAN BGP EVPN fabric. EVPN Multi-Site provisioning for Layer-2/Layer-3 DCI across multiple fabrics using the MSD fabric, is documented in a separate chapter. The deployment details of how overlay Networks and VRFs can be easily provisioned from the DCNM, is covered under Creating and Deploying Networks and VRFs.
Guidelines for VXLAN BGP EVPN Fabrics Provisioning
-
For any switch to be successfully imported into DCNM, the user specified for discovery/import, should have the following permissions:
-
SSH access to the switch
-
Ability to perform SNMPv3 queries
-
Ability to run the show commands including show run, show interfaces, etc.
-
-
The switch discovery user need not have the ability to make any configuration changes on the switches. It is primarily used for read access.
-
When an invalid command is deployed by DCNM to a device, for example, a command with an invalid key chain due to an invalid entry in the fabric settings, an error is generated displaying this issue. This error is not cleared after correcting the invalid fabric entry. You need to manually cleanup or delete the invalid commands to clear the error.
Note that the fabric errors related to the command execution are automatically cleared only when the same failed command succeeds in the subsequent deployment.
-
LAN credentials are required to be set of any user that needs to be perform any write access to the device. LAN credentials need to be set on the DCNM, on a per user per device basis. When a user imports a device into the Easy Fabric, and LAN credentials are not set for that device, DCNM moves this device to a migration mode. Once the user sets the appropriate LAN credentials for that device, a subsequent Save & Deploy will retrigger the device import process.
-
The Save & Deploy button triggers the intent regeneration for the entire fabric as well as a configuration compliance check for all the switches within the fabric. This button is required but not limited to the following cases:
-
A switch or a link is added, or any change in the topology
-
A change in the fabric settings that must be shared across the fabric
-
A switch is removed or deleted
-
A new vPC pairing or unpairing is done
-
A change in the role for a device
When you click Save & Deploy, the changes in the fabric are evaluated, and the configuration for the entire fabric is generated. You can preview the generated configuration, and then deploy it at a fabric level. Therefore, Save & Deploy can take more time depending on the size of the fabric.
When you right-click on a switch icon, you can use the Deploy Config option to deploy per switch configurations. This option is a local operation for a switch, that is, the expected configuration or intent for a switch is evaluated against it’s current running configuration, and a config compliance check is performed for the switch to get the In-Sync or Out-of-Sync status. If the switch is out of sync, the user is provided with a preview of all the configurations running in that particular switch that vary from the intent defined by the user for that respective switch.
-
-
Persistent configuration diff is seen for the command line: system nve infra-vlan int force . The persistent diff occurs if you have deployed this command via the freeform configuration to the switch. Although the switch requires the force keyword during deployment, the running configuration that is obtained from the switch in DCNM does not display the force keyword. Therefore, the system nve infra-vlan int force command always shows up as a diff.
The intent in DCNM contains the line:
system nve infra-vlan int force
The running config contains the line:
system nve infra-vlan int
As a workaround to fix the persistent diff, edit the freeform config to remove the force keyword after the first deployment such that it is system nve infra-vlan int .
The force keyword is required for the initial deploy and must be removed after a successful deploy. You can confirm the diff by using the Side-by-side Comparison tab in the Config Preview window.
The persistent diff is also seen after a write erase and reload of a switch. Update the intent on DCNM to include the force keyword, and then you need to remove the force keyword after the first deployment.
-
When the switch contains the hardware access-list tcam region arp-ether 256 command, which is deprecated without the double-wide keyword, the below warning is displayed:
WARNING: Configuring the arp-ether region without "double-wide" is deprecated and can result in silent non-vxlan packet drops. Use the "double-wide" keyword when carving TCAM space for the arp-ether region.
Since the original hardware access-list tcam region arp-ether 256 command does not match the policies in DCNM, this config is captured in the switch_freeform policy. After the hardware access-list tcam region arp-ether 256 double-wide command is pushed to the switch, the original tcam command that does not contain the double-wide keyword is removed.
You must manually remove the hardware access-list tcam region arp-ether 256 command from the switch_freeform policy. Otherwise, config compliance shows a persistent diff.
Here is an example of the hardware access-list command on the switch:
switch(config)# show run | inc arp-ether switch(config)# hardware access-list tcam region arp-ether 256 Warning: Please save config and reload the system for the configuration to take effect switch(config)# show run | inc arp-ether hardware access-list tcam region arp-ether 256 switch(config)# switch(config)# hardware access-list tcam region arp-ether 256 double-wide Warning: Please save config and reload the system for the configuration to take effect switch(config)# show run | inc arp-ether hardware access-list tcam region arp-ether 256 double-wide
You can see that the original tcam command is overwritten.
Creating a New VXLAN BGP EVPN Fabric
This procedure shows how to create a new VXLAN BGP EVPN fabric.
This procedure contains descriptions for the IPv4 underlay. For information about IPv6 underlay, see IPv6 Underlay Support for Easy Fabric.
-
Choose Control > Fabric Builder.
The Fabric Builder window appears. When you log in for the first time, the Fabrics section has no entries. After you create a fabric, it is displayed on the Fabric Builder window, wherein a rectangular box represents each fabric.
A standalone or member fabric contains Switch_Fabric (in the Type field), the AS number (in the ASN field), and mode of replication (in the Replication Mode field).
-
Click Create Fabric, the Add Fabric screen appears.
The fields are explained:
Fabric Name - Enter the name of the fabric.
Fabric Template - From the drop-down menu, choose the Easy_Fabric_11_1 fabric template. The fabric settings for creating a standalone fabric appear.
The tabs and their fields in the screen are explained in the subsequent points. The overlay and underlay network parameters are included in these tabs.
Note
If you are creating a standalone fabric as a potential member fabric of an MSD fabric (used for provisioning overlay networks for fabrics that are connected through EVPN Multi-Site technology), then browse through the Multi-Site Domain for VXLAN BGP EVPN Fabrics topic before member fabric creation.
-
The General tab is displayed by default. The fields in this tab are:
BGP ASN: Enter the BGP AS number the fabric is associated with.
Enable IPv6 Underlay: Enable the IPv6 underlay feature. For information, see IPv6 Underlay Support for Easy Fabric.
Enable IPv6 Link-Local Address: Enables the IPv6 Link-Local address.
Fabric Interface Numbering : Specifies whether you want to use point-to-point (p2p) or unnumbered networks.
Underlay Subnet IP Mask - Specifies the subnet mask for the fabric interface IP addresses.
Underlay Routing Protocol : The IGP used in the fabric, OSPF, or IS-IS.
Route-Reflectors (RRs) – The number of spine switches that are used as route reflectors for transporting BGP traffic. Choose 2 or 4 from the drop-down box. The default value is 2.
To deploy spine devices as RRs, DCNM sorts the spine devices based on their serial numbers, and designates two or four spine devices as RRs. If you add more spine devices, existing RR configuration will not change.
Increasing the count - You can increase the route reflectors from two to four at any point in time. Configurations are automatically generated on the other two spine devices designated as RRs.
Decreasing the count - When you reduce four route reflectors to two, remove the unneeded route reflector devices from the fabric. Follow these steps to reduce the count from 4 to 2.
-
Change the value in the drop-down box to 2.
-
Identify the spine switches designated as route reflectors.
An instance of the rr_state policy is applied on the spine switch if it is a route reflector. To find out if the policy is applied on the switch, right-click the switch, and choose View/edit policies. In the View/Edit Policies screen, search rr_state in the Template field. It is displayed on the screen.
-
Delete the unneeded spine devices from the fabric (right-click the spine switch icon and choose Discovery > Remove from fabric).
If you delete existing RR devices, the next available spine switch is selected as the replacement RR.
-
Click Save & Deploy in the fabric topology window.
You can preselect RRs and RPs before performing the first Save & Deploy operation. For more information, see Preselecting Switches as Route-Reflectors and Rendezvous-Points.
Anycast Gateway MAC : Specifies the anycast gateway MAC address.
NX-OS Software Image Version : Select an image from the list.
If you upload Cisco NX-OS software images through the image upload option, the uploaded images are listed in this field. If you select an image, and save the Fabric Settings, the system checks that all the switches within the fabric have the selected version. If some devices do not run the image, a warning is prompted to perform an In-Service Software Upgrade (ISSU) to the specified image. The warning is also accompanied with a Resolve button. This takes the user to the image management screen with the mismatched switches auto selected for device upgrade/downgrade to the specified NX-OS image specified in Fabric Settings. Till, all devices run the specified image, the deployment process is incomplete.
If you want to deploy more than one type of software image on the fabric switches, don’t specify any image. If an image is specified, delete it.
-
-
Click the Replication tab. Most of the fields are auto generated. You can update the fields if needed.
Replication Mode : The mode of replication that is used in the fabric for BUM (Broadcast, Unknown Unicast, Multicast) traffic. The choices are Ingress Replication or Multicast. When you choose Ingress replication, the multicast related fields get disabled.
You can change the fabric setting from one mode to the other, if no overlay profile exists for the fabric.
Multicast Group Subnet : IP address prefix used for multicast communication. A unique IP address is allocated from this group for each overlay network.
In the DCNM 11.0(1) release, the replication mode change is not allowed if a policy template instance is created for the current mode. For example, if a multicast related policy is created and deployed, you cannot change the mode to Ingress.
Enable Tenant Routed Multicast (TRM) – Select the check box to enable Tenant Routed Multicast (TRM) that allows overlay multicast traffic to be supported over EVPN/MVPN in the VXLAN BGP EVPN fabric.
Default MDT Address for TRM VRFs: The multicast address for Tenant Routed Multicast traffic is populated. By default, this address is from the IP prefix specified in the Multicast Group Subnet field. When you update either field, ensure that the TRM address is chosen from the IP prefix specified in Multicast Group Subnet.
For more information, see Overview of Tenant Routed Multicast.
Rendezvous-Points - Enter the number of spine switches acting as rendezvous points.
RP mode – Choose from the two supported multicast modes of replication, ASM (for Any-Source Multicast [ASM]) or BiDir (for Bidirectional PIM [BIDIR-PIM]).
When you choose ASM, the BiDir related fields are not enabled. When you choose BiDir, the BiDir related fields are enabled.
Note
BIDIR-PIM is supported on Cisco's Cloud Scale Family platforms 9300-EX and 9300-FX/FX2, and software release 9.2(1) onwards.
When you create a new VRF for the fabric overlay, this address is populated in the Underlay Multicast Address field, in the Advanced tab.
Underlay RP Loopback ID – The loopback ID used for the rendezvous point (RP), for multicast protocol peering purposes in the fabric underlay.
The next two fields are enabled if you choose BIDIR-PIM as the multicast mode of replication.
Underlay Primary RP Loopback ID – The primary loopback ID used for the phantom RP, for multicast protocol peering purposes in the fabric underlay.
Underlay Backup RP Loopback ID – The secondary loopback ID used for the phantom RP, for multicast protocol peering purposes in the fabric underlay.
Underlay Second Backup RP Loopback Id and Underlay Third Backup RP Loopback Id: Used for the second and third fallback Bidir-PIM Phantom RP.
-
Click the vPC tab. Most of the fields are auto generated. You can update the fields if needed.
vPC Peer Link VLAN – VLAN used for the vPC peer link SVI.
Make vPC Peer Link VLAN as Native VLAN - Enables vPC peer link VLAN as Native VLAN.
vPC Peer Keep Alive option – Choose the management or loopback option. If you want to use IP addresses assigned to the management port and the management VRF, choose management. If you use IP addresses assigned to loopback interfaces (and a non-management VRF), choose loopback.
If you use IPv6 addresses, you must use loopback IDs.
vPC Auto Recovery Time - Specifies the vPC auto recovery time-out period in seconds.
vPC Delay Restore Time - Specifies the vPC delay restore period in seconds.
vPC Peer Link Port Channel ID - Specifies the Port Channel ID for a vPC Peer Link. By default, the value in this field is 500.
vPC IPv6 ND Synchronize – Enables IPv6 Neighbor Discovery synchronization between vPC switches. The check box is enabled by default. Clear the check box to disable the function.
vPC advertise-pip - Select the check box to enable the Advertise PIP feature.
You can enable the advertise PIP feature on a specific vPC as well. For more information, see Advertising PIP on vPC.
Enable the same vPC Domain Id for all vPC Pairs: Enable the same vPC Domain ID for all vPC pairs. When you select this field, the vPC Domain Id field is editable.
vPC Domain Id - Specifies the vPC domain ID to be used on all vPC pairs.
vPC Domain Id Range - Specifies the vPC Domain Id range to use for new pairings.
Enable Qos for Fabric vPC-Peering - Enable QoS on spines for guaranteed delivery of vPC Fabric Peering communication. For more information, see QoS for Fabric vPC-Peering.
Note
QoS for vPC fabric peering and queuing policies options in fabric settings are mutually exclusive.
Qos Policy Name - Specifies QoS policy name that should be same on all fabric vPC peering spines. The default name is spine_qos_for_fabric_vpc_peering.
-
Click the Protocols tab. Most of the fields are auto generated. You can update the fields if needed.
Underlay Routing Loopback Id - The loopback interface ID is populated as 0 since loopback0 is usually used for fabric underlay IGP peering purposes.
Underlay VTEP Loopback Id - The loopback interface ID is populated as 1 since loopback1 is used for the VTEP peering purposes.
Underlay Routing Protocol Tag - The tag defining the type of network.
OSPF Area ID – The OSPF area ID, if OSPF is used as the IGP within the fabric.
Note
The OSPF or IS-IS authentication fields are enabled based on your selection in the Underlay Routing Protocol field in the General tab.
Enable OSPF Authentication – Select the check box to enable OSPF authentication. Deselect the check box to disable it. If you enable this field, the OSPF Authentication Key ID and OSPF Authentication Key fields get enabled.
OSPF Authentication Key ID - The Key ID is populated.
OSPF Authentication Key - The OSPF authentication key must be the 3DES key from the switch.
Note
Plain text passwords are not supported. Log in to the switch, retrieve the encrypted key and enter it in this field. Refer, Retrieving the Authentication Key section for details.
IS-IS Level - Select the IS-IS level from this drop-down list.
Enable IS-IS Authentication - Select the check box to enable IS-IS authentication. Deselect the check box to disable it. If you enable this field, the IS-IS authentication fields are enabled.
IS-IS Authentication Keychain Name - Enter the Keychain name, such as CiscoisisAuth.
IS-IS Authentication Key ID - The Key ID is populated.
IS-IS Authentication Key - Enter the Cisco Type 7 encrypted key.
Note
Plain text passwords are not supported. Log in to the switch, retrieve the encrypted key and enter it in this field. Refer the Retrieving the Authentication Key section for details.
Enable BGP Authentication - Select the check box to enable BGP authentication. Deselect the check box to disable it. If you enable this field, the BGP Authentication Key Encryption Type and BGP Authentication Key fields are enabled.
Note
If you enable BGP authentication using this field, leave the iBGP Peer-Template Config field blank to avoid duplicate configuration.
BGP Authentication Key Encryption Type – Choose the 3 for 3DES encryption type, or 7 for Cisco encryption type.
BGP Authentication Key - Enter the encrypted key based on the encryption type.
Note
Plain text passwords are not supported. Log in to the switch, retrieve the encrypted key and enter it in the BGP Authentication Key field. Refer the Retrieving the Authentication Key section for details.
Enable PIM Hello Authentication - Enables the PIM hello authentication.
PIM Hello Authentication Key - Specifies the PIM hello authentication key.
Enable BFD: Select the check box to enable feature bfd on all switches in the fabric. This feature is valid only on IPv4 underlay and the scope is within a fabric.
From Cisco DCNM Release 11.3(1), BFD within a fabric is supported natively. The BFD feature is disabled by default in the Fabric Settings. If enabled, BFD is enabled for the underlay protocols with the default settings. Any custom required BFD configurations must be deployed via the per switch freeform or per interface freeform policies.
The following config is pushed after you select the Enable BFD check box:
feature bfd
Note
After you upgrade from DCNM Release 11.2(1) with BFD enabled to DCNM Release 11.3(1), the following configurations are pushed on all P2P fabric interfaces:
no ip redirects no ipv6 redirects
For information about BFD feature compatibility, refer your respective platform documentation and for information about the supported software images, see Compatibility Matrix for Cisco DCNM.
Enable BFD for iBGP: Select the check box to enable BFD for the iBGP neighbor. This option is disabled by default.
Enable BFD for OSPF: Select the check box to enable BFD for the OSPF underlay instance. This option is disabled by default, and it is grayed out if the link state protocol is ISIS.
Enable BFD for ISIS: Select the check box to enable BFD for the ISIS underlay instance. This option is disabled by default, and it is grayed out if the link state protocol is OSPF.
Enable BFD for PIM: Select the check box to enable BFD for PIM. This option is disabled by default, and it is be grayed out if the replication mode is Ingress.
Here are the examples of the BFD global policies:
router ospf <ospf tag> bfd router isis <isis tag> address-family ipv4 unicast bfd ip pim bfd router bgp <bgp asn> neighbor <neighbor ip> bfd
Enable BFD Authentication: Select the check box to enable BFD authentication. If you enable this field, the BFD Authentication Key ID and BFD Authentication Key fields are editable.
Note
BFD Authentication is not supported when the Fabric Interface Numbering field under the General tab is set to unnumbered. The BFD authentication fields will be grayed out automatically. BFD authentication is valid for only for P2P interfaces.
BFD Authentication Key ID: Specifies the BFD authentication key ID for the interface authentication. The default value is 100.
BFD Authentication Key: Specifies the BFD authentication key.
For information about how to retrieve the BFD authentication parameters, see Retrieving the Encrypted BFD Authentication Key.
iBGP Peer-Template Config – Add iBGP peer template configurations on the leaf switches to establish an iBGP session between the leaf switch and route reflector.
If you use BGP templates, add the authentication configuration within the template and clear the Enable BGP Authentication check box to avoid duplicate configuration.
In the sample configuration, the 3DES password is displayed after password 3.
router bgp 65000 password 3 sd8478fswerdfw3434fsw4f4w34sdsd8478fswerdfw3434fsw4f4w
Until Cisco DCNM Release 11.3(1), iBGP peer template for iBGP definition on the leafs or border role devices and BGP RRs were same. From DCNM Release 11.4(1), the following fields can be used to specify different configurations:
-
iBGP Peer-Template Config – Specifies the config used for RR and spines with border role.
-
Leaf/Border/Border Gateway iBGP Peer-Template Config – Specifies the config used for leaf, border, or border gateway. If this field is empty, the peer template defined in iBGP Peer-Template Config is used on all BGP enabled devices (RRs, leafs, border, or border gateway roles).
In brownfield migration, if the spine and leaf use different peer template names, both iBGP Peer-Template Config and Leaf/Border/Border Gateway iBGP Peer-Template Config fields need to be set according to the switch config. If spine and leaf use the same peer template name and content (except for the “route-reflector-client” CLI), only iBGP Peer-Template Config field in fabric setting needs to be set. If the fabric settings on iBGP peer templates do not match the existing switch configuration, an error message is generated and the migration will not proceed.
-
-
Click the Advanced tab. Most of the fields are auto generated. You can update the fields if needed.
VRF Template and VRF Extension Template: Specifies the VRF template for creating VRFs, and the VRF extension template for enabling VRF extension to other fabrics.
Network Template and Network Extension Template: Specifies the network template for creating networks, and the network extension template for extending networks to other fabrics.
Site ID - The ID for this fabric if you are moving this fabric within an MSD. The site ID is mandatory for a member fabric to be a part of an MSD. Each member fabric of an MSD has a unique site ID for identification.
Intra Fabric Interface MTU - Specifies the MTU for the intra fabric interface. This value should be an even number.
Layer 2 Host Interface MTU - Specifies the MTU for the layer 2 host interface. This value should be an even number.
Power Supply Mode - Choose the appropriate power supply mode.
CoPP Profile - Choose the appropriate Control Plane Policing (CoPP) profile policy for the fabric. By default, the strict option is populated.
VTEP HoldDown Time - Specifies the NVE source interface hold down time.
Brownfield Overlay Network Name Format: Enter the format to be used to build the overlay network name during a brownfield import or migration. The network name should not contain any white spaces or special characters except underscore (_) and hyphen (-). The network name must not be changed once the brownfield migration has been initiated. See the Creating Networks for the Standalone Fabric section for the naming convention of the network name. The syntax is [<string> | $$VLAN_ID$$] $$VNI$$ [<string>| $$VLAN_ID$$] and the default value is Auto_Net_VNI$$VNI$$_VLAN$$VLAN_ID$$. When you create networks, the name is generated according to the syntax you specify. The following table describes the variables in the syntax.
Variables
Description
$$VNI$$
Specifies the network VNI ID found in the switch configuration. This is a mandatory keyword required to create unique network names.
$$VLAN_ID$$
Specifies the VLAN ID associated with the network.
VLAN ID is specific to switches, hence DCNM picks the VLAN ID from one of the switches, where the network is found, randomly and use it in the name.
We recommend not to use this unless the VLAN ID is consistent across the fabric for the VNI.
<string>
This variable is optional and you can enter any number of alphanumeric characters that meet the network name guidelines.
Example overlay network name: Site_VNI12345_VLAN1234
Note
Ignore this field for greenfield deployments. The Brownfield Overlay Network Name Format applies for the following brownfield imports:
-
CLI-based overlays
-
Configuration profile-based overlay where the configuration profiles were created in Cisco DCNM Release
10.4(2).
Enable CDP for Bootstrapped Switch - Enables CDP on management (mgmt0) interface for bootstrapped switch. By default, for bootstrapped switches, CDP is disabled on the mgmt0 interface.
Enable VXLAN OAM - Enables the VXLAM OAM functionality for devices in the fabric. This is enabled by default. Clear the check box to disable VXLAN OAM function.
If you want to enable the VXLAN OAM function on specific switches and disable on other switches in the fabric, you can use freeform configurations to enable OAM and disable OAM in the fabric settings.
Note
The VXLAN OAM feature in Cisco DCNM is only supported on a single fabric or site.
Enable Tenant DHCP – Select the check box to enable feature dhcp and associated configurations globally on all switches in the fabric. This is a pre-requisite for support of DHCP for overlay networks that are part of the tenant VRFs.
Note
Ensure that Enable Tenant DHCP is enabled before enabling DHCP related parameters in the overlay profiles.
Enable NX-API - Specifies enabling of NX-API on HTTPS. This check box is checked by default.
Enable NX-API on HTTP on Port - Specifies enabling of NX-API on HTTP. Enable this check box and the Enable NX-API check box to use HTTP. This check box is checked by default. If you uncheck this check box, the applications that use NX-API and supported by Cisco DCNM, such as Endpoint Locator (EPL), Layer 4-Layer 7 services (L4-L7 services), VXLAN OAM, and so on, start using the HTTPS instead of HTTP.
Note
If you check the Enable NX-API check box and the Enable NX-API on HTTP check box, applications use HTTP.
Enable Policy-Based Routing (PBR) - Select this check box to enable routing of packets based on the specified policy. Starting with Cisco NX-OS Release 7.0(3)I7(1) and later releases, this feature works on Cisco Nexus 9000 Series switches with Nexus 9000 Cloud Scale (Tahoe) ASICs. This feature is used along with the Layer 4-Layer 7 service workflow. For information on Layer 4-Layer 7 service, refer the Layer 4-Layer 7 Service chapter.
Enable Strict Config Compliance - Enable the Strict Config Compliance feature by selecting this check box. By default, this feature is disabled. For more information, refer Strict Configuration Compliance.
Enable AAA IP Authorization - Enables AAA IP authorization, when IP Authorization is enabled in the remote authentication server. This is required to support DCNM in scenarios where customers have strict control of which IP addresses can have access to the switches.
Enable DCNM as Trap Host - Select this check box to enable DCNM as a SNMP trap destination. Typically, for a native HA DCNM deployment, the eth1 VIP IP address will be configured as SNMP trap destination on the switches. By default, this check box is enabled.
Greenfield Cleanup Option – Enable the switch cleanup option for switches imported into DCNM with Preserve-Config=No, without a switch reload. This option is typically recommended only for the fabric environments with Cisco Nexus 9000v Switches to improve on the switch clean up time. The recommended option for Greenfield deployment is to employ Bootstrap or switch cleanup with a reboot. In other words, this option should be unchecked.
Enable Precision Time Protocol (PTP): Enables PTP across a fabric. When you select this check box, PTP is enabled globally and on core-facing interfaces. Additionally, the PTP Source Loopback Id and PTP Domain Id fields are editable. For more information, see Precision Time Protocol for Easy Fabric.
PTP Source Loopback Id: Specifies the loopback interface ID Loopback that is used as the Source IP Address for all PTP packets. The valid values range from 0 to 1023. The PTP loopback ID cannot be the same as RP, Phantom RP, NVE, or MPLS loopback ID. Otherwise, an error will be generated. The PTP loopback ID can be the same as BGP loopback or user-defined loopback which is created from DCNM.
If the PTP loopback ID is not found during Save & Deploy, the following error is generated:
Loopback interface to use for PTP source IP is not found. Create PTP loopback interface on all the devices to enable PTP feature.
PTP Domain Id: Specifies the PTP domain ID on a single network. The valid values range from 0 to 127.
Enable MPLS Handoff: Select the check box to enable the MPLS Handoff feature. For more information, see the Border Provisioning Use Case in VXLAN BGP EVPN Fabrics - MPLS SR and LDP Handoff chapter.
Underlay MPLS Loopback Id: Specifies the underlay MPLS loopback ID. The default value is 101.
Enable TCAM Allocation: TCAM commands are automatically generated for VXLAN and vPC Fabric Peering when enabled.
Enable Default Queuing Policies: Check this check box to apply QoS policies on all the switches in this fabric. To remove the QoS policies that you applied on all the switches, uncheck this check box, update all the configurations to remove the references to the policies, and save and deploy. From Cisco DCNM Release 11.3(1), pre-defined QoS configurations are included that can be used for various Cisco Nexus 9000 Series Switches. When you check this check box, the appropriate QoS configurations are pushed to the switches in the fabric. The system queuing is updated when configurations are deployed to the switches. You can perform the interface marking with defined queuing policies, if required, by adding the required configuration to the per interface freeform block.
From Cisco DCNM Release 11.4(1), the DSCP mapping for QoS 5 has changed from 40 to 46 in the policy template. For DCNM 11.3(1) deployments that have been upgraded to 11.4(1), you will see the diffs that need to be deployed.
Review the actual queuing policies by opening the policy file in the template editor. From Cisco DCNM Web UI, choose Control > Template Library. Search for the queuing policies by the policy file name, for example, queuing_policy_default_8q_cloudscale. Choose the file and click the Modify/View template icon to edit the policy.
See the Cisco Nexus 9000 Series NX-OS Quality of Service Configuration Guide for platform specific details.
N9K Cloud Scale Platform Queuing Policy: Choose the queuing policy from the drop-down list to be applied to all Cisco Nexus 9200 Series Switches and the Cisco Nexus 9000 Series Switches that ends with EX, FX, and FX2 in the fabric. The valid values are queuing_policy_default_4q_cloudscale and queuing_policy_default_8q_cloudscale. Use the queuing_policy_default_4q_cloudscale policy for FEXes. You can change from the queuing_policy_default_4q_cloudscale policy to the queuing_policy_default_8q_cloudscale policy only when FEXes are offline.
N9K R-Series Platform Queuing Policy: Choose the queuing policy from the drop-down list to be applied to all Cisco Nexus switches that ends with R in the fabric. The valid value is queuing_policy_default_r_series.
Other N9K Platform Queuing Policy: Choose the queuing policy from the drop-down list to be applied to all other switches in the fabric other than the switches mentioned in the above two options. The valid value is queuing_policy_default_other.
Enable MACsec - Enables MACsec for the fabric. For more information, see MACsec Support in Easy Fabric and eBGP Fabric.
Freeform CLIs - Fabric level freeform CLIs can be added while creating or editing a fabric. They are applicable to switches across the fabric. You must add the configurations as displayed in the running configuration, without indentation. Switch level freeform configurations such as VLAN, SVI, and interface configurations should only be added on the switch. For more information, refer Enabling Freeform Configurations on Fabric Switches.
Leaf Freeform Config - Add CLIs that should be added to switches that have the Leaf, Border, and Border Gateway roles.
Spine Freeform Config - Add CLIs that should be added to switches with a Spine, Border Spine, Border Gateway Spine, and Super Spine roles.
Intra-fabric Links Additional Config - Add CLIs that should be added to the intra-fabric links.
-
-
Click the Resources tab.
Manual Underlay IP Address Allocation – Do not select this check box if you are transitioning your VXLAN fabric management to DCNM.
-
By default, DCNM allocates the underlay IP address resources (for loopbacks, fabric interfaces, etc) dynamically from the defined pools. If you select the check box, the allocation scheme switches to static, and some of the dynamic IP address range fields are disabled.
-
For static allocation, the underlay IP address resources must be populated into the Resource Manager (RM) using REST APIs.
Refer the Cisco DCNM REST API Reference Guide, Release 11.2(1) for more details. The REST APIs must be invoked after the switches are added to the fabric, and before you use the Save & Deploy option.
-
The Underlay RP Loopback IP Range field stays enabled if BIDIR-PIM function is chosen for multicast replication.
-
Changing from static to dynamic allocation keeps the current IP resource usage intact. Only future IP address allocation requests are taken from dynamic pools.
Underlay Routing Loopback IP Range - Specifies loopback IP addresses for the protocol peering.
Underlay VTEP Loopback IP Range - Specifies loopback IP addresses for VTEPs.
Underlay RP Loopback IP Range - Specifies the anycast or phantom RP IP address range.
Underlay Subnet IP Range - IP addresses for underlay P2P routing traffic between interfaces.
Underlay MPLS Loopback IP Range: Specifies the underlay MPLS loopback IP address range.
For eBGP between Border of Easy A and Easy B, Underlay routing loopback and Underlay MPLS loopback IP range must be a unique range. It should not overlap with IP ranges of the other fabrics, else VPNv4 peering will not come up.
Layer 2 VXLAN VNI Range and Layer 3 VXLAN VNI Range - Specifies the VXLAN VNI IDs for the fabric.
Network VLAN Range and VRF VLAN Range - VLAN ranges for the Layer 3 VRF and overlay network.
Subinterface Dot1q Range - Specifies the subinterface range when L3 sub interfaces are used.
VRF Lite Deployment - Specify the VRF Lite method for extending inter fabric connections.
The VRF Lite Subnet IP Range field specifies resources reserved for IP address used for VRF LITE when VRF LITE IFCs are auto-created. If you select Back2BackOnly, ToExternalOnly, or Back2Back&ToExternal then VRF LITE IFCs are auto-created.
Auto Deploy Both - This check box is applicable for the symmetric VRF Lite deployment. When you select this check box, it would set the auto deploy flag to true for auto-created IFCs to turn on symmetric VRF Lite configuration.
This check box can be selected or deselected when the VRF Lite Deployment field is not set to Manual. In the case, a user explicitly unchecks the auto-deploy field for any auto-created IFCs, then the user input is always given the priority. This flag only affects the new auto-created IFC and it does not affect the existing IFCs.
VRF Lite Subnet IP Range and VRF Lite Subnet Mask – These fields are populated with the DCI subnet details. Update the fields as needed.
The values shown in your screen are automatically generated. If you want to update the IP address ranges, VXLAN Layer 2/Layer 3 network ID ranges or the VRF/Network VLAN ranges, ensure the following:
Note
When you update a range of values, ensure that it does not overlap with other ranges. You should only update one range of values at a time. If you want to update more than one range of values, do it in separate instances. For example, if you want to update L2 and L3 ranges, you should do the following.
-
Update the L2 range and click Save.
-
Click the Edit Fabric option again, update the L3 range and click Save.
Service Network VLAN Range - Specifies a VLAN range in the Service Network VLAN Range field. This is a per switch overlay service network VLAN range. The minimum allowed value is 2 and the maximum allowed value is 3967.
Route Map Sequence Number Range - Specifies the route map sequence number range. The minimum allowed value is 1 and the maximum allowed value is 65534.
-
-
Click the Manageability tab.
The fields in this tab are:
DNS Server IPs - Specifies the comma separated list of IP addresses (v4/v6) of the DNS servers.
DNS Server VRFs - Specifies one VRF for all DNS servers or a comma separated list of VRFs, one per DNS server.
NTP Server IPs - Specifies comma separated list of IP addresses (v4/v6) of the NTP server.
NTP Server VRFs - Specifies one VRF for all NTP servers or a comma separated list of VRFs, one per NTP server.
Syslog Server IPs – Specifies the comma separated list of IP addresses (v4/v6) IP address of the syslog servers, if used.
Syslog Server Severity – Specifies the comma separated list of syslog severity values, one per syslog server. The minimum value is 0 and the maximum value is 7. To specify a higher severity, enter a higher number.
Syslog Server VRFs – Specifies one VRF for all syslog servers or a comma separated list of VRFs, one per syslog server.
AAA Freeform Config – Specifies the AAA freeform configurations.
If AAA configurations are specified in the fabric settings, switch_freeform PTI with source as UNDERLAY_AAA and description as AAA Configurations will be created.
- Click the Bootstrap tab.
Enable Bootstrap - Select this check box to enable the bootstrap feature. Bootstrap allows easy day-0 import and bring-up of new devices into an existing fabric. Bootstrap leverages the NX-OS POAP functionality.
After you enable bootstrap, you can enable the DHCP server for automatic IP address assignment using one of the following methods:
-
External DHCP Server: Enter information about the external DHCP server in the Switch Mgmt Default Gateway and Switch Mgmt IP Subnet Prefix fields.
-
Local DHCP Server: Enable the Local DHCP Server check box and enter details for the remaining mandatory fields.
Enable Local DHCP Server - Select this check box to initiate enabling of automatic IP address assignment through the local DHCP server. When you select this check box, the DHCP Scope Start Address and DHCP Scope End Address fields become editable.
If you do not select this check box, DCNM uses the remote or external DHCP server for automatic IP address assignment.
DHCP Version – Select DHCPv4 or DHCPv6 from this drop-down list. When you select DHCPv4, the Switch Mgmt IPv6 Subnet Prefix field is disabled. If you select DHCPv6, the Switch Mgmt IP Subnet Prefix is disabled.
Note
Cisco Nexus 9000 and 3000 Series Switches support IPv6 POAP only when switches are either Layer-2 adjacent (eth1 or out-of-band subnet must be a /64) or they are L3 adjacent residing in some IPv6 /64 subnet. Subnet prefixes other than /64 are not supported.
DHCP Scope Start Address and DHCP Scope End Address - Specifies the first and last IP addresses of the IP address range to be used for the switch out of band POAP.
Switch Mgmt Default Gateway - Specifies the default gateway for the management VRF on the switch.
Switch Mgmt IP Subnet Prefix - Specifies the prefix for the Mgmt0 interface on the switch. The prefix should be between 8 and 30.
DHCP scope and management default gateway IP address specification - If you specify the management default gateway IP address 10.0.1.1 and subnet mask 24, ensure that the DHCP scope is within the specified subnet, between 10.0.1.2 and 10.0.1.254.
Switch Mgmt IPv6 Subnet Prefix - Specifies the IPv6 prefix for the Mgmt0 interface on the switch. The prefix should be between 112 and 126. This field is editable if you enable IPv6 for DHCP.
Enable AAA Config – Select this check box to include AAA configurations from the Manageability tab as part of the device startup config post bootstrap.
Bootstrap Freeform Config - (Optional) Enter additional commands as needed. For example, if you require some additional configurations to be pushed to the device and be available post device bootstrap, they can be captured in this field, to save the desired intent. After the devices boot up, they will contain the configuration defined in the Bootstrap Freeform Config field.
Copy-paste the running-config to a freeform config field with correct indentation, as seen in the running configuration on the NX-OS switches. The freeform config must match the running config. For more information, see Resolving Freeform Config Errors in Switches.
DHCPv4/DHCPv6 Multi Subnet Scope - Specifies the field to enter one subnet scope per line. This field is editable after you check the Enable Local DHCP Server check box.
The format of the scope should be defined as:
DHCP Scope Start Address, DHCP Scope End Address, Switch Management Default Gateway, Switch Management Subnet Prefix
For example: 10.6.0.2, 10.6.0.9, 10.6.0.1, 24
-
-
Click the Configuration Backup tab. The fields on this tab are:
Hourly Fabric Backup: Select the check box to enable an hourly backup of fabric configurations and the intent.
The hourly backups are triggered during the first 10 minutes of the hour.
Scheduled Fabric Backup: Check the check box to enable a daily backup. This backup tracks changes in running configurations on the fabric devices that are not tracked by configuration compliance.
Scheduled Time: Specify the scheduled backup time in a 24-hour format. This field is enabled if you check the Scheduled Fabric Backup check box.
Select both the check boxes to enable both back up processes.
The backup process is initiated after you click Save.
The scheduled backups are triggered exactly at the time you specify with a delay of up to two minutes. The scheduled backups are triggered regardless of the configuration deployment status.
The backup configuration files are stored in the following path in DCNM: /usr/local/cisco/dcm/dcnm/data/archive
The number of archived files that can be retained is set in the # Number of archived files per device to be retained: field in the Server Properties window.
Note
To trigger an immediate backup, do the following:
-
Choose Control > Fabric Builder. The Fabric Builder screen comes up.
-
Click within the specific fabric box. The fabric topology screen comes up.
-
From the Actions pane at the left part of the screen, click Re-Sync Fabric.
You can also initiate the fabric backup in the fabric topology window. Click Backup Now in the Actions pane.
-
-
Click ThousandEyes Agent tab. This feature is supported on Cisco DCNM Release 11.5(3) only. For more information, refer to Configuring Global Settings for ThousandEyes Enterprise Agent.
The fields on this tab are:
Note
The fabric settings for ThousandEyes Agent overwrites the global settings and applies the same configuration for all the ThousandEyes Agent installed on switches in that fabric.
-
Enable Fabric Override for ThousandEyes Agent Installation: Select the check box to enable the ThousandEyes Enterprise Agent on the fabric.
-
ThousandEyes Account Group Token: Specifies ThousandEyes Enterprise Agent account group token for installation.
-
VRF on Switch for ThousandEyes Agent Collector Reachability: Specifies the VRF data which provides internet reachability.
-
DNS Domain: Specifies the switch DNS domain configuration.
-
DNS Server IPs: Specifies the comma separated list of IP addresses (v4/v6) of Domain Name System (DNS) server. You can enter a maximum of three IP addresses for the DNS Server.
-
NTP Server IPs: Specifies comma separated list of IP addresses (v4/v6) of Network Time Protocol (NTP) server. You can enter a maximum of three IP addresses for the NTP Server.
-
Enable Proxy for Internet Access: Select the check box to enable the proxy setting for NX-OS switch internet access.
-
Proxy Information: Specifies the proxy server port information.
-
Proxy Bypass: Specifies the server list for which proxy is bypassed.
-
-
Click Save after filling and updating relevant information. A note appears briefly at the bottom right part of the screen, indicating that the fabric is created. When a fabric is created, the fabric page comes up. The fabric name appears at the top left part of the screen.
(At the same time, the newly created fabric instance appears on the Fabric Builder screen. To go to the Fabric Builder screen, click the left arrow (←) button above the Actions pane [to the left of the screen]).
The Actions pane allows you to perform various functions. One of them is the Add switches option to add switches to the fabric. After you create a fabric, you should add fabric devices. The options are explained:
-
Tabular View - By default, the switches are displayed in the topology view. Use this option to view switches in the tabular view.
-
Refresh topology - Allows you to refresh the topology.
-
Save Layout – Saves a custom view of the topology. You can create a specific view in the topology and save it for ease of use.
-
Delete saved layout – Deletes the custom view of the topology
-
Topology views - You can choose between Hierarchical, Random and Custom saved layout display options.
-
Hierarchical - Provides an architectural view of your topology. Various Switch Roles can be defined that draws the nodes on how you configure your CLOS topology.
-
Random - Nodes are placed randomly on the window. DCNM tries to make a guess and intelligently place nodes that belong together in close proximity.
-
Custom saved layout - You can drag nodes around to your liking. Once you have the positions as how you like, you can click Save Layout to remember the positions. Next time you come to the topology, DCNM will draw the nodes based on your last saved layout positions.
-
-
Restore Fabric – Allows you to restore the fabric to a prior DCNM configuration state (one month back, two months back, and so on). For more information, see the Restore Fabric section.
-
Backup Now: You can initiate a fabric backup manually by clicking Backup Now. Enter a name for the tag and click OK. Regardless of the settings you choose under the Configuration Backup tab in the Fabric Settings dialog box, you can initiate a backup using this option.
-
Resync Fabric - Use this option to resynchronize DCNM state when there is a large scale out-of-band change, or if configuration changes do not register in the DCNM properly. The resync operation does a full CC run for the fabric switches and recollects “show run” and “show run all” commands from the switches. When you initiate the re-sync process, a progress message is displayed on the window. During the re-sync, the running configuration is taken from the switches. Then, the Out-of-Sync/In-Sync status for the switch is recalculated based on the intent or expected configuration defined in DCNM versus the current running configuration that was taken from the switches.
-
Add Switches – Allows you to add switch instances to the fabric.
-
Fabric Settings – Allows you to view or edit fabric settings.
-
Cloud icon - Click the Cloud icon to display (or not display) an Undiscovered cloud.
When you click the icon, the Undiscovered cloud and its links to the selected fabric topology are not displayed.
Click the Cloud icon again to display the Undiscovered cloud.
-
SCOPE - You can toggle between fabrics by using the SCOPE drop-down box at the top right. The current fabric is highlighted. An MSD and its member fabrics are distinctly displayed, wherein the member fabrics are indented, under the MSD fabric.
Adding Switches to a Fabric
Switches in each fabric are unique, and hence, each switch can only be added to one fabric.
Click the Add Switches option from the Actions panel to add switches to the fabric created in DCNM. The Inventory Management screen comes up. The screen contains two tabs, one for discovering existing switches and the other for discovering new switches. Both options are explained.
Additionally, you can pre-provision switches and interfaces. For more information, see Pre-provisioning a Device and Pre-provisioning an Ethernet Interface.
Note |
When DCNM discovers a switch with the hostname containing the period character (.), it is treated as a domain-name and truncated. Only the text prior to the period character (.) is considered as a hostname. For example:
|
Discovering Existing Switches
-
After clicking on Add Switches, use the Discover Existing Switches tab to add one or more existing switches into the fabric. In this case, a switch with known credentials and a pre-provisioned IP address, is added to the fabric. The IP address (Seed IP), administrator username, and password (Username and Password fields) of the switch are provided as the input by a user. The Preserve Config knob is set to yes by default. This is the option that a user would select for a brownfield import of a device into the fabric. For a greenfield import where the device configuration will be cleaned up as part of the import process, the user should set the Preserve Config knob to no.
Note
Easy_Fabric_eBGP does not support brownfield import of a device into the fabric.
-
Click Start discovery. The Scan Details window comes up shortly. Since the Max Hops field was populated with 2 (by default), the switch with the specified IP address (leaf-91) and switches two hops from that switch, are populated in the Scan Detailsresult.
-
If the DCNM was able to perform a successful shallow discovery to a switch, the status will show up as Manageable. Select the check box next to the appropriate switch(es) and click Import into fabric.
Though this example describes the discovery of one switch, multiple switches can be discovered at once.
The switch discovery process is initiated. The Progress column displays progress for all the selected switches. It displays done for each switch on completion.
Note
You must not close the screen (and try to add switches again) until all selected switches are imported or an error message comes up.
If an error message comes up, close the screen. The fabric topology screen comes up. The error messages are displayed at the top right part of the screen. Resolve the errors wherever applicable and initiate the import process again by clicking Add Switches in the Actions panel.
DCNM discovers all the switches, and the Progress column displays done for all switches, close the screen. The Standalone fabric topology screen comes up again. The switch icons of the added switches are displayed in it.
Note
You will encounter the following errors during switch discovery sometimes.
-
Click Refresh topology to view the latest topology view.
When all switches are added and roles assigned to them, the fabric topology contains the switches and connections between them.
-
After discovering the devices, assign an appropriate role to each device. For this purpose, tight click the device, and use the Set role option to set the appropriate role. Alternatively, the tabular view may be employed to assign the same role to multiple devices at one go.
If you choose the Hierarchical layout for display (in the Actions panel), the topology automatically gets aligned as per role assignment, with the leaf devices at the bottom, the spine devices connected on top of them, and the border devices at the top.
Assign vPC switch role - To designate a pair of switches as a vPC switch pair, right-click the switch and choose the vPC peer switch from the list of switches.
AAA server password - During fabric creation, if you have entered AAA server information (in the Manageability tab), you must update the AAA server password on each switch. Else, switch discovery fails.
When a new vPC pair is created and deployed successfully using Cisco DCNM, one of the peers might be out-of-sync for the no ip redirects CLI even if the command exists on the switch. This out-of-sync is due to a delay on the switch to display the CLI in the running configuration, which causes a diff in the configuration compliance. Re-sync the switches in the Config Deployment window to resolve the diff.
-
Click Save & Deploy at the top right part of the screen.
The template and interface configurations form the underlay network configuration on the switches. Also, freeform CLIs that were entered as part of fabric settings (leaf and spine switch freeform configurations entered in the Advanced tab) are deployed. For more details on freeform configurations, refer Enabling Freeform Configurations on Fabric Switches.
Configuration Compliance: If the provisioned configurations and switch configurations do not match, the Status column displays out-of-sync. For example, if you enable a function on the switch manually through a CLI, then it results in a configuration mismatch.
To ensure configurations provisioned from DCNM to the fabric are accurate or to detect any deviations (such as out-of-band changes), DCNM’s Configuration Compliance engine reports and provides necessary remediation configurations.
When you click Save & Deploy, the Config Deployment window appears.
If the status is out-of-sync, it suggests that there is inconsistency between the DCNM and configuration on the device.
The Re-sync button is displayed for each switch in the Re-sync column. Use this option to resynchronize DCNM state when there is a large scale out-of-band change, or if configuration changes do not register in the DCNM properly. The re-sync operation does a full CC run for the switch and recollects “show run” and “show run all” commands from the switch. When you initiate the re-sync process, a progress message is displayed on the screen. During the re-sync, the running configuration is taken from the switch. The Out-of-Sync/In-Sync status for the switch is recalculated based on the intent defined in DCNM.
Click the Preview Config column entry (updated with a specific number of lines). The Config Preview screen comes up.
The PendingConfig tab displays the pending configurations for successful deployment.
The Side-by-sideComparison tab displays the current configurations and expected configurations together.
In DCNM 11, multi-line banner motd configuration is supported. Multi-line banner motd configuration can be configured in DCNM with freeform configuration policy, either per switch using switch_freeform, or per fabric using leaf/spine freeform configuration. Note that after the multi-line banner motd is configured, deploy the policy by executing the Save & Deploy option in the (top right part of the) fabric topology screen. Else, the policy may not be deployed properly on the switch. The banner policy is only to configure single-line banner configuration. Also, you can only create one banner related freeform configuration/policy. Multiple policies for configuring banner motd are not supported.
-
Close the screen.
In the Configuration Deployment screen, click Deploy Config at the bottom part of the screen to initiate pending configuration onto the switch. The Status column displays FAILED or SUCCESS state. For a FAILED status, investigate the reason for failure to address the issue.
After successful configuration provisioning (when all switches display a progress of 100%), close the screen.
The fabric topology is displayed. The switch icons turn green to indicate successful configuration.
If a switch icon is in red color, it indicates that the switch and DCNM configurations are not in sync.When deployment is pending on a switch, the switch is displayed in blue color. The pending state indicates that there is a pending deployment or pending recomputation. You can click on the switch and review the pending deployments using Preview or Deploy Config options, or click Save & Deploy to recompute the state of the switch.
Note
If there are any warning or errors in the CLI execution, a notification will appear in the Fabric builder window. Warnings or errors that are auto-resolvable have the Resolve option.
When a leaf switch boots up after a switch reload or RMA operation, DCNM provisions configurations for the switch and FEX devices connected to it. Occasionally, FEX connectivity comes up after DCNM provisions FEX (host interface) configurations, resulting in a configuration mismatch. To resolve the mismatch, click Save & Deploy again in the fabric topology screen.
From Cisco NX-OS Release 11.4(1), if you uncheck the FEX check box in the Topology window, FEX devices are hidden in the Fabric Builder topology window as well. To view FEX in Fabric Builder, you need to check this check box. This option is applicable for all fabrics and it is saved per session or until you log out of DCNM. If you log out and log in to DCNM, the FEX option is reset to default, that is, enabled by default. For more information, see Show Panel.
An example of the Deploy Config option usage is for switch-level freeform configurations. Refer Enabling Freeform Configurations on Fabric Switches for details.
Discovering New Switches
-
When a new Cisco NX-OS device is powered on, typically that device has no startup configuration or any configuration state for that matter. Consequently, it powers on with NX-OS and post initialization, goes into a POAP loop. The device starts sending out DHCP requests on all the interfaces that are up including the mgmt0 interface.
-
As long as there is IP reachability between the device and the DCNM, the DHCP request from the device, will be forwarded to the DCNM. For easy day-0 device bring-up, the bootstrap options should be enabled in the Fabric Settings as mentioned earlier.
-
With bootstrap enabled for the fabric, the DHCP request coming from the device will be serviced by the DCNM. The temporary IP address allocated to the device by the DCNM will be employed to learn basic information about the switch including the device model, device NX-OS version, etc.
-
In the DCNM GUI, go to a fabric (Click Control > Fabric Builder and click a fabric). The fabric topology is displayed.
Go to the fabric topology window and click the Add switches option from the Actions panel. The Inventory Management window comes up.
-
Click the POAP tab.
As mentioned earlier, DCNM retrieves the serial number, model number, and version from the device and displays them on the Inventory Management along window. Also, an option to add the IP address, hostname, and password are made available. If the switch information is not retrieved, refresh the window.
Note
-
At the top left part of the window, export and import options are provided to export and import the .csv file that contains the switch information. You can pre-provision devices using the import option as well.
Select the checkbox next to the switch and enter the switch credentials: IP address and host name.
Based on the IP address of your device, you can either add the IPv4 or IPv6 address in the IP Address field.
Beginning with Release 11.2(1), you can provision devices in advance. To pre-provision devices, refer to Pre-provisioning a Device.
-
-
In the Admin Password and Confirm Admin Password fields, enter and confirm the admin password.
This admin password is applicable for all the switches displayed in the POAP window.
Note
If you do not want to use admin credentials to discover switches, you can instead use the AAA authentication, that is, RADIUS or TACACS credentials for discovery only.
-
(Optional) Use discovery credentials for discovering switches.
-
Click the Add Discovery Credentials icon to enter the discovery credentials for switches.
-
In the Discovery Credentials window, enter the discovery credentials such as discovery username and password.
Click OK to save the discovery credentials.
If the discovery credentials are not provided, DCNM uses the admin user and password to discover switches.
-
-
Click Bootstrap at the top right part of the screen.
DCNM provisions the management IP address and other credentials to the switch. In this simplified POAP process, all ports are opened up.
-
Click Refresh Topology to get updated information. The added switch goes through the POAP cycle. Monitor and check the switch for POAP completion.
-
After the added switch completes POAP, the fabric builder topology page is refreshed with the added switch thereby depicting its discovered physical connections. Set the appropriate role for the switch followed by a Save & Deploy operation at the fabric level. The Fabric Settings, switch role, the topology etc. are evaluated by the Fabric Builder and the appropriate intended configuration for the switch is generated as part of the Save operation. The pending configuration will provide a list of the configurations that need to be deployed to the new switch in order to bring it IN-SYNC with the intent.
Note
For any changes on the fabric that results in the Out-of-Sync, then you must deploy the changes. The process is the same as explained in the Discovering Existing Switches section.
During fabric creation, if you have entered AAA server information (in the Manageability tab), you must update the AAA server password on each switch. Else, switch discovery fails.
-
After the pending configurations are deployed, the Progress column displays 100% for all switches.
-
Click Close to return to the fabric builder topology.
-
Click Refresh Topology to view the update. All switches must be in green color indicating that they are functional.
-
The switch and the link are discovered in DCNM. Configurations are built based on various policies (such as fabric, topology, and switch generated policies). The switch image (and other required) configurations are enabled on the switch.
-
In the DCNM GUI, the discovered switches can be seen in the Standalone fabric topology. Up to this step, the POAP is completed with basic settings. You must setup interfaces through the Control > Interfaces option for any additional configurations, but not limited to the following:
-
vPC pairing.
-
Breakout interfaces.
-
Port channels, and adding members to ports.
When you enable or disable a vPC pairing/un-pairing or the advertise-pip option, or update Multi-Site configuration, you should use the Save & Deploy operation. At the end of the operation, an error prompts you to configure the shutdown or no shutdown command on the nve interface. A sample error screenshot when you enable a vPC setup:
To resolve, go to the Control > Interfaces screen and deploy the Shutdown operation on the nve interface followed by a No Shutdown configuration. This is depicted in the figure below where the up arrow corresponds to a No Shutdown operation while a down arrow corresponds to a Shutdown operation.
-
You can right-click the switch to view various options:
-
Set Role - Assign a role to the switch (Spine, Border Gateway, and so on).
Note
-
Changing of the switch role is allowed only before executing Save & Deploy.
-
Starting from DCNM 11.1(1), switch roles can be changed if there are no overlays on the switches, but only as per the list of allowed switch role changes given at Switch Operations.
-
-
Modes - Maintenance and Active/Operational modes.
-
vPC Pairing - Select a switch for vPC and then select its peer.
You can create a virtual link for a vPC pair or change the existing physical link to a virtual link for a vPC pair.
-
Manage Interfaces - Deploy configurations on the switch interfaces.
-
View/Edit Policies - See switch policies and edit them as required.
-
History - View per switch deployment and policy change history.
The Policy Change History tab lists the history of policies along with the users who made the changes like add, update, or delete.
Under the Policy Change History tab, for a policy, click Detailed History under the Generated Config column to view the generated config before and after.
The following table provides the summary of generated config before and after for Policy Template Instances (PTIs).
PTI Operations Generated Config Before Generated Config After Add Empty Contains the config Update Contains config before changes Contains config after changes Mark-Delete Contains the config to be removed. Contains the config to be removed with colour change. Delete Contains the config Empty Note
When a policy or profile template is applied, an instance is created for each application of the template, which is known as Policy Template Instance or PTI.
-
Preview Config - View the pending configuration and the side-by-side comparison of the running and expected configuration.
-
Deploy Config - Deploy per switch configurations.
-
Discovery - You can use this option to update the credentials of the switch, reload the switch, rediscover the switch, and remove the switch from the fabric.
The new fabric is created, the fabric switches are discovered in DCNM, the underlay configuration provisioned on those switches, and the configurations between DCNM and the switches are synced. The remaining tasks are:
-
Provision interface configurations such as vPCs, loopback interface, and subinterface configurations. [Refer Interfaces].
-
Create networks and deploy them on the switches. [Refer Creating and Deploying Networks and VRFs].
Pre-provisioning Support in DCNM 11
Cisco DCNM supports provisioning of device configuration in advance. This is specifically applicable for scenarios where devices have been procured, but not yet delivered or received by the Customers. The purchase order typically has information about the device serial number, device model and so on, which in turn can be used to prepare the device configuration in DCNM prior to the device connectivity to the Network. Pre-provisioning is supported for Cisco NX-OS devices in both Easy Fabric and External/Classic_LAN fabrics.
Pre-provisioning a Device
Note |
Ensure that you enter DHCP details in the Bootstrap tab in the fabric settings. |
-
The pre-provisioned devices support the following configurations in DCNM:
-
Base management
-
vPC Pairing
-
Intra-Fabric links
-
Ethernet ports
-
Port-channel
-
vPC
-
ST FEX
-
AA FEX
-
Loopback
-
Overlay network configurations
-
-
The pre-provisioned devices do not support the following configurations in DCNM:
-
Inter-Fabric links
-
Sub-interface
-
Interface breakout configuration
-
-
When a device is being pre-provisioned has breakout links, you need to specify the corresponding breakout command along with the switch's model and gateway in the Data field in the Add a new device to pre-provisioning window in order to generate the breakout PTI.
Note the following guidelines:
-
Multiple breakout commands can be separated by a semicolon (;).
-
The definitions of the fields in the data JSON object are as follows:
-
modulesModel: (Mandatory) Specifies the switch module’s model information.
-
gateway: (Mandatory) Specifies the default gateway for the management VRF on the switch. This field is required to create the intent to pre-provision devices. You must enter the gateway even if it is in the same subnet as DCNM to create the intent as part of pre-provisioning a device.
-
breakout: (Optional) Specifies the breakout command provided in the switch.
-
portMode: (Optional) Specifies the port mode of the breakout interface.
-
The examples of the values in the Data field are as follows:
-
{"modulesModel": ["N9K-C93180LC-EX"], "gateway": "10.1.1.1/24"}
-
{"modulesModel": ["N9K-C93180LC-EX"],"breakout": "interface breakout module 1 port 1 map 10g-4x", "portMode": "hardware profile portmode 4x100G+28x40G", "gateway": "172.22.31.1/24" }
-
{"modulesModel": ["N9K-X9736C-EX", "N9K-X9732C-FX", "N9K-C9516-FM-E2", "N9K-C9516-FM-E2", "N9K-C9516-FM-E2", "N9K-C9516-FM-E2", "N9K-SUP-B+", "N9K-SC-A", "N9K-SC-A"], "gateway": "172.22.31.1/24"}
-
{"breakout":"interface breakout module 1 port 50 map 10g-4x" , "gateway": "172.16.1.1/24", "modulesModel": ["N9K-C93180YC-EX "]}
-
{"modulesModel": ["N9K-X9732C-EX", "N9K-X9732C-EX", "N9K-C9504-FM-E", "N9K-C9504-FM-E", "N9K-SUP-B", "N9K-SC-A", "N9K-SC-A"], "gateway": "172.29.171.1/24", "breakout":"interface breakout module 1 port 1,11,19 map 10g-4x; interface breakout module 1 port 7 map 25g-4x"}
-
{"modulesModel": ["N9K-C93180LC-EX"], "gateway": "10.1.1.1/24","breakout": "interface breakout module 1 port 1-4 map 10g-4x","portMode": "hardware profile portmode 48x25G + 2x100G + 4x40G"}
-
Procedure
Step 1 |
Click Control > Fabric Builder. The Fabric Builder screen is displayed. |
||
Step 2 |
Click within the fabric box. |
||
Step 3 |
From the Actions panel, click the Add switches option. The Inventory Management screen is displayed. |
||
Step 4 |
Click the POAP tab. |
||
Step 5 |
In the POAP tab, do the following:
IP Address: Specify the IPv4 or IPv6 address of the new device. Serial Number: The serial number for the new device. Serial number is found in the Cisco Build of Material Purchase and you can refer to these values while using the pre-provisioning feature. For information about the Data field, see the examples provided in guidelines. The device details appear in the POAP screen. You can add more devices for pre-provisioning. At the top left part of the window, Export and Import icons are provided to export and import the .csv file that contains the switch information. Using the Import option, you can pre-provision multiple devices. Add new devices’ information in the .csv file with all the mandatory fields (SerialNumber, Model, version, IpAddress, Hostname, and Data fields [JSON Object]). The Data column consists of the model name of the module to identify the hardware type from the fabric template. A .csv file screenshot: |
||
Step 6 |
Enter the administration password in the Admin Password and Confirm Admin Password fields. |
||
Step 7 |
Select the device(s) and click Bootstrap at the top right part of the screen. The leaf1 device appears in the fabric topology. From the Actions panel, click Tabular View. You cannot deploy the fabric till the status of all the pre-provisioned switch(es) are displayed as ok under the Discovery Status column.
When you connect leaf1 to the fabric, the switch is provisioned with the IP address 10.1.1.1. |
||
Step 8 |
Navigate to Fabric Builder and set roles for the device. Create intra-link policy using one of the templates:
Click Save & Deploy. Configuration for the switches are captured in corresponding PTIs and can be seen in the View/Edit Policies window. |
||
Step 9 |
To bring in the physical device, you can follow the manual RMA or POAP RMA procedure. For more information, see Return Material Authorization (RMA). If you use the POAP RMA procedure, ignore the error message of failing to put the device into maintenance mode due to no connectivity since it is expected to have no connectivity to a non-existing device. You need to click Save & Deploy in the fabric after one or more switches are online to provision the host ports. This action must be performed before overlays are provisioned for the host port attachment. |
Pre-provisioning an Ethernet Interface
From DCNM Release 11.4(1), you can pre-provision Ethernet interfaces in the Interface window. This pre-provisioning feature is supported in the Easy, External, and eBGP fabrics. You can add Ethernet interfaces to only pre-provisioned devices before they are discovered in DCNM.
Note |
Before attaching a network/VRF, you must pre-provision the Ethernet interface before adding it to Port-channels, vPCs, ST FEX, AA FEX, loopback, subinterface, tunnel, ethernet, and SVI configurations. |
Before you begin
Procedure
Step 1 |
Navigate to the fabric containing the pre-provisioned device from the Fabric Builder window. |
||
Step 2 |
Right click the pre-provisioned device and select Manage Interfaces. You can also navigate to the Interfaces window by selecting Control > Fabrics > Interfaces. From the Scope drop-down list, select the fabric containing the pre-provisioned device. |
||
Step 3 |
Click Add. |
||
Step 4 |
Enter all the required details in the Add Interface window. Type: Select Ethernet from this drop-down list. Select a device: Select the pre-provisioned device.
Enter Interface Name: Enter a valid interface name based on the module type. For example, Ethernet1/1, eth1/1, or e1/1. The interface with same name should be available on the device after it is added. Policy: Select a policy that should be applied on the interface. For more information, see Adding Interfaces. |
||
Step 5 |
Click Save. |
||
Step 6 |
Click Preview to check the expected configuration that will be deployed to the switch after it is added.
|
Pre-provisioning a vPC Pair
Before you begin
Ensure that you have enabled Bootstrap in the Fabric Settings.
Procedure
Step 1 |
Import both the devices into the fabric. For instructions, see Pre-provisioning a Device. The following example in the image shows two Cisco Nexus 9000 Series devices that are pre-provisioned and added to an existing Fabric. Choose Add Switches in the Action panel. On the Inventory Management screen, click PowerOn Auto Provisioning (POAP). The devices will show up in the fabric as gray/undiscovered devices. |
Step 2 |
Right click and select appropriate roles for these devices similar to other reachable devices. |
Step 3 |
To create vPC pairing between the devices with physical peer-link or MCT, perform the following steps: |
Pre-provisioning a vPC Host Interface
Procedure
Step 1 |
Create physical ethernet interfaces on the pre-provisioned devices. Add a vPC host interface similar to a regular vPC pair or switches. For instructions, see Pre-provisioning an Ethernet Interface. For example, leaf1-leaf2 represents the pre-provisioned vPC device pair, assuming that Ethernet interfaces 1/1 is already pre-provisioned on both devices leaf1 and leaf2. |
Step 2 |
Create a vPC host truck interface as shown in the following image. Preview and Deploy actions doesn't yield any result, because both require the device to be present. The vPC host interface is created and displays status as Not discovered as shown in the following image. |
Attaching Overlays to Pre-provisioned Devices
Overlay VRFs and Networks can be attached to pre-provisioned devices similar to any other discovered device.
The following example shows where an overlay network is attached to the pre-provisioned vPC pair of leafs (leaf1-leaf2). It is also attached to the pre-provisioned vPC host interface port-channels created on leaf1-leaf2.
Preview and Deploy operations are disabled for the pre-provisioned devices, because the devices are not reachable. After the pre-provisioned device is reachable, all operations are enabled similar to other discovered devices.
On Fabric Builder > View/Edit Policies, you can view the entire intent generated for the pre-provisioned device, including the overlay network/VRF attachment information as shown in the following image.
Precision Time Protocol for Easy Fabric
In the fabric settings for the Easy_Fabric_11_1 template, select the Enable Precision Time Protocol (PTP) check box to enable PTP across a fabric. When you select this check box, PTP is enabled globally and on core-facing interfaces. Additionally, the PTP Loopback Id and PTP Domain Id fields are editable.
The PTP feature works only when all the devices in a fabric are cloud-scale devices. Warnings are displayed if there are non-cloud scale devices in the fabric, and PTP is not enabled. Examples of the cloud-scale devices are Cisco Nexus 93180YC-EX, Cisco Nexus 93180YC-FX, Cisco Nexus 93240YC-FX2, and Cisco Nexus 93360YC-FX2 switches.
For more information, see the Configuring PTP chapter in Cisco Nexus 9000 Series NX-OS System Management Configuration Guide and Cisco Network Insights for Resources Application for Cisco DCNM User Guide.
For LAN fabric deployments, specifically in a VXLAN EVPN based fabric deployments, you have to enable PTP globally, and also enable PTP on core-facing interfaces. The interfaces could be configured to the external PTP server like a VM or Linux-based machine. Therefore, the interface should be edited to have a connection with the grandmaster clock.
It is recommended that the grandmaster clock should be configured outside of Easy Fabric and it is IP reachable. The interfaces toward the grandmaster clock need to be enabled with PTP via the interface freeform config.
All core-facing interfaces are auto-enabled with the PTP configuration after you click Save & Deploy. This action ensures that all devices are PTP synced to the grandmaster clock. Additionally, for any interfaces that are not core-facing, such as interfaces on the border devices and leafs that are connected to hosts, firewalls, service-nodes, or other routers, the ttag related CLI must be added. The ttag is added for all traffic entering the VXLAN EVPN fabric and the ttag must be stripped when traffic is exiting this fabric.
Here is the sample PTP configuration:
feature ptp
ptp source 100.100.100.10 -> IP address of the loopback interface (loopback0) that is already created or user created loopback interface in the fabric settings
ptp domain 1 -> PTP domain ID specified in fabric settings
interface Ethernet1/59 -> Core facing interface
ptp
interface Ethernet1/50 -> Host facing interface
ttag
ttag-strip
The following guidelines are applicable for PTP:
-
The PTP feature can be enabled in a fabric when all the switches in the fabric have Cisco NX-OS Release 7.0(3)I7(1) or a higher version. Otherwise, the following error message is displayed:
PTP feature can be enabled in the fabric, when all the switches have NX-OS Release 7.0(3)I7(1) or higher version. Please upgrade switches to NX-OS Release 7.0(3)I7(1) or higher version to enable PTP in this fabric.
-
For hardware telemetry support in NIR, the PTP configuration is a prerequisite.
-
If you are adding a non-cloud scale device to an existing fabric which contains PTP configuration, the following warning is displayed:
TTAG is enabled fabric wide, when all devices are cloud scale switches so it cannot be enabled for newly added non cloud scale device(s).
-
If a fabric contains both cloud scale and non-cloud scale devices, the following warning is displayed when you try to enable PTP:
TTAG is enabled fabric wide, when all devices are cloud scale switches and is not enabled due to non cloud scale device(s).
Support for Super Spine Role in DCNM
Super Spine is a device that is used for interconnecting multiple spine-leaf PODs. Prior to the DCNM Release 11.3(1), it was possible to interconnect multiple VXLAN EVPN Easy fabrics via super spines. However, these super spines had to be part of an external fabric. Within each Easy Fabric, an appropriate IGP is used for underlay connectivity. eBGP between the super spine layer in the external fabric and spine layer in the Easy Fabrics would be the recommended way of interconnecting multiple VXLAN EVPN Easy Fabrics. The eBGP peering can be configured via inter-fabric links or an appropriate mix of interface and eBGP configuration on the respective switches.
From DCNM Release 11.3(1), you have an extra interconnectivity option with super spines. You can have multiple spine-leaf PODs within the same Easy Fabric that are interconnected via super spines such that the same IGP domain extends across all the PODs, including the super spines. Within such a deployment, the BGP RRs and RPs (if applicable) are provisioned on the super spine layer. The spine layer becomes a pseudo interconnect between the leafs and super spines. VTEPs may be optionally hosted on the super spines if they have the border functionality.
The following Super Spine roles are supported in DCNM:
-
Super Spine
-
Border Super Spine
-
Border Gateway Super Spine
A border super spine handles multiple functionalities including the functionalities of a super spine, RR, RP (optionally), and a border leaf. Similarly, a border gateway super spine serves a super spine, RR, RP (optional), and a border gateway. It’s not recommended to overload border functionality on the super spine or RR layer. Instead, attach border leafs or border gateways to the super spine layer for external connectivity. The super spine layer serves as the interconnect with the RR or RP functionality.
The following are the characteristics of super spine switch roles in DCNM:
-
Supported only for the Easy_Fabric_11_1 template.
-
Can only connect to spines and borders. The valid connections are:
-
Spines to super spines
-
Spines to border super spines and border gateway super spines
-
Super spines to border leafs and border gateway leafs
-
-
RR or RP (if applicable) functionality is always be configured on super spines if they are present in a fabric. The maximum number of 4 RRs and RPs are supported even with Super Spines.
-
Border Super Spine and Border Gateway Super Spine roles are supported for inter-fabric connections.
-
vPC configurations aren’t supported on super spines.
-
Super spines don’t support IPv6 underlay configuration.
-
During the Brownfield import of switches, if a switch has the super spine role, the following error is displayed:
Serial number: [super spine/border super spine/border gateway superspine] Role isn’t supported with preserved configuration yes option.
Supported Topologies for Super Spine Switches
DCNM supports the following topologies with super spine switches.
Topology 1: Super Spine Switches in a Spine Leaf Topology
In this topology, leaf switches are connected to spines, and spines are then connected to Super Spines switches which can be super spines, border super spines, border gateway super spines.
Topology 2: Super Spine Switches Connected to Border
In this topology, there are four leaf switches connecting to the Spine switches, which are connected to the two Super Spine switches. These Super Spine switches are connected to the border or border gateway leaf switches.
Adding a Super Spine Switch to an Existing VXLAN BGP EVPN Fabric
Procedure
Step 1 |
Navigate to Control > Fabric Builder. |
||
Step 2 |
From the Fabric Builder window, click Add Switches in the actions panel. For more information, see Adding Switches to a Fabric. |
||
Step 3 |
Right-click an existing switch or the newly added switch, and use the Set role option to set the appropriate super spine role.
|
||
Step 4 |
Click Save & Deploy. An error is displayed saying: Adding new switch with Super Spine role is not allowed, if save&deploy has already been performed in the fabric without any super spine role switch. |
||
Step 5 |
Click the error, and click the Resolve button. A confirmation dialog box is displayed asking whether you want to continue. If you click Yes, the following actions are performed by DCNM:
You should not add a device(s) with super spine, border super spine, or border gateway super spine role if the same will be connected to a border spine or border gateway spine that is already present in the fabric. This action will result in the below error after clicking Save & Deploy. If you want to use the existing device(s) with border spine roles, you need to remove the same and add them again with the appropriate role (spine or super spine and its variants) and valid connections. |
Changing the TCAM Configuration on a Device
If you are onboarding the Cisco Nexus 9300 Series switches and Cisco Nexus 9500 Series switches with X9500 line cards using the bootstrap feature with POAP, DCNM pushes the following policies depending on the switch models:
-
Cisco Nexus 9300 Series Switches: tcam_pre_config_9300 and tcam_pre_config_vxlan
-
Cisco Nexus 9500 Series Switches: tcam_pre_config_9500 and tcam_pre_config_vxlan
Perform the following steps to change the TCAM carving of a device in DCNM.
-
Choose Control > Fabrics > Fabric Builder.
-
Click the fabric containing the specified switches that have been onboarded using the bootstrap feature.
-
Click Tabular View under the Actions menu in the Fabric Builder window.
-
Select all the specified switches and click the View/Edit Policies icon.
-
Search for tcam_pre_config policies.
-
If the TCAM config is incorrect or not applicable, select all these policies and click the Delete icon to delete policies.
-
Add one or multiple tcam_config policies and provide the correct TCAM configuration. For more information about how to add a policy, see Adding PTIs for Multiple Switches.
-
Reload the respective switches.
If the switch is used as a leaf, border leaf, border gateway leaf, border spine, or border gateway spine, add the tcam_config policy with the following command and deploy.
hardware access-list tcam region racl 1024
This config is required on the switches so that the NGOAM and VXLAN Suppress ARP features are functional.
Make sure that the priority of this tcam_config policy is higher than the tcam_pre_config_vxlan policy so that the config policy with racl 1024 is configured before the tcam_pre_config_vxlan policy.
Note |
The tcam_pre_config_vxlan policy contains the config: hardware access-list tcam region arp-ether 256 double-wide. |
Preselecting Switches as Route-Reflectors and Rendezvous-Points
This task shows how to preselect switches as Route-Reflectors (RRs) and Rendezvous-Points (RPs) before the first Save & Deploy operation.
Note |
This scenario is applicable when you have more than 2 spines and you want to control the preselection of RRs and RPs before the first Save & Deploy operation. |
Procedure
Step 1 |
Import switches successfully. |
||
Step 2 |
Create the rr_state or rp_state policies using View/Edit Policies on the spines or super spine switches, which should be preselected as RR or RP.
|
||
Step 3 |
Click Save & Deploy, and then click Deploy Config. The spines that have rr_state policies become RR and spines that have rp_state policies become RP. |
||
Step 4 |
After Save & Deploy, if you want to replace the preselected RRs and RPs with a new set of devices, then old RR and RP devices should be removed from the fabric before performing the same steps. |
Adding a vPC L3 Peer Keep-Alive Link
This procedure shows how to add a vPC L3 peer keep-alive link.
Note |
|
Procedure
Step 1 |
From DCNM, navigate to Control > Template Library. |
Step 2 |
Search for the vpc_serial_simulated policy, select it, and click the Edit icon. |
Step 3 |
Edit the template properties and set the Template Sub Type to Device so that this policy appears in View/Edit Policies. |
Step 4 |
Navigate to the Fabric Builder window and click on the fabric containing the vPC pair switches. |
Step 5 |
Click Tabular View and select the vPC pair switches, and then click View/Edit Policies. You can also right-click the switches individually in the topology and select View/Edit Policies. |
Step 6 |
Click + to add policies. |
Step 7 |
From the Policy drop-down list, select vpc_serial_simulated policy and add priority. Click Save. Note that if both switches are selected, then this policy will be created on both vPC pair switches. |
Step 8 |
Navigate back to Tabular View and click the Links tab. |
Step 9 |
Select the link between vPC pair, which has to be a vPC peer keep alive and click Edit. |
Step 10 |
From the Link Template drop-down list, select int_intra_vpc_peer_keep_alive_link_11_1. Enter values for the remaining fields. Make sure to leave the field empty for the default VRF and click Save. |
Step 11 |
Click Save & Deploy, and click Preview Config for one of the switches. If VRF is non-default, use switch_freeform to create the respective VRF. Navigate to the topology and click the vPC pair switch to see the details. |
Changing the Local Authentication to AAA Authentication for Switches in a Fabric
Procedure
Step 1 |
Log in to DCNM and navigate to Control > Fabric Builder. |
Step 2 |
Click the Edit icon for a fabric and add the AAA authentication commands in the AAA Freeform Config field under the Manageability tab. |
Step 3 |
In the Fabric Builder topology window, click Add Switches. Use the AAA credentials in this window to add switches into the DCNM. |
Step 4 |
If you are importing switches in to the fabric via POAP, you need to have the AAA configs on the switch. Navigate to the fabric settings and add the relevant commands in Bootstrap Freeform Config. |
Step 5 |
In the Fabric Builder topology window, click Add Switches. In the PowerON Auto Provisioning (POAP) tab, click the Add discovery credentials icon and enter the discovery credentials. Click Save & Deploy after you complete adding switches. |
IPv6 Underlay Support for Easy Fabric
From Cisco DCNM Release 11.3(1), you can create an Easy fabric with IPv6 only underlay. The IPv6 underlay is supported only for the Easy_Fabric_11_1 template. For more information, see Configuring a VXLAN Fabric with IPv6 Underlay.
Brownfield Deployment-Transitioning VXLAN Fabric Management to DCNM
DCNM supports Brownfield deployments, wherein you transition your VXLAN BGP EVPN fabric management to DCNM. The transition involves migrating existing network configurations to DCNM. For information, see Managing a Brownfield VXLAN BGP EVPN Fabric.
Configuring Fabrics with eBGP Underlay
You can use the Easy_Fabric_eBGP fabric template to create a fabric with eBGP underlay. For more information, see Managing BGP-Based Routed Fabrics and Managing a Greenfield VXLAN BGP EVPN Fabric.
Creating an External Fabric
In DCNM 11.1(1) release, you can add switches to the external fabric. Generic pointers:
-
An external fabric is a monitor-only or managed mode fabric. DCNM supports only the monitor mode for Cisco IOS-XR family devices.
-
You can import, remove, and delete switches for an external fabric.
-
For Inter-Fabric Connection (IFC) cases, you can choose Cisco 9000, 7000 and 5600 Series switches as destination switches in the external fabric.
-
You can use non-existing switches as destination switches.
-
The template that supports an external fabric is External_Fabric.
-
If an external fabric is an MSD fabric member, then the MSD topology screen displays the external fabric with its devices, along with the member fabrics and their devices.
When viewed from an external fabric topology screen, any connections to non-DCNM managed switches are represented by a cloud icon labeled as Undiscovered.
-
You can set up a Multi-Site or a VRF-lite IFC by manually configuring the links for the border devices in the VXLAN fabric or by using an automatic Deploy Border Gateway Method or VRF Lite IFC Deploy Method. If you are configuring the links manually for the border devices, we recommend using the Core Router role to set up a Multi-Site eBGP underlay from a Border Gateway device to a Core Router and the Edge Router role to set up a VRF-lite Inter-Fabric Connection (IFC) from a Border device to an Edge device.
-
If you are using the Cisco Nexus 7000 Series Switch with Cisco NX-OS Release 6.2(24a) on the LAN Classic or External fabrics, make sure to enable AAA IP Authorization in the fabric settings.
-
You can discover the following non-Nexus devices in an external fabric:
-
IOS-XE family devices: Cisco CSR 1000v, Cisco IOS XE Gibraltar 16.10.x, Cisco ASR 1000 Series routers, and Cisco Catalyst 9000 Series Switches
-
IOS-XR family devices: ASR 9000 Series Routers, IOS XR Release 6.5.2 and Cisco NCS 5500 Series Routers, IOS XR Release 6.5.3
-
Arista 4.2 (Any model)
-
-
Configure all the non-Nexus devices, except Cisco CSR 1000v, before adding them to the external fabric.
-
From Cisco DCNM Release 11.4(1), you can configure non-Nexus devices as borders. You can create an IFC between a non-Nexus device in an external fabric and a Cisco Nexus device in an easy fabric. The interfaces supported for these devices are:
-
Routed
-
Subinterface
-
Loopback
-
-
From Cisco DCNM, Release 11.4(1), you can configure a Cisco ASR 1000 Series routers and Cisco Catalyst 9000 Series switches as edge routers, set up a VRF-lite IFC and connect it as a border device with an easy fabric.
-
Before a VDC reload, discover Admin VDC in the fabric. Otherwise, the reload operation does not occur.
-
You can connect a Cisco data center to a public cloud using Cisco CSR 1000v. See the Connecting Cisco Data Center and a Public Cloud chapter for a use case.
-
In an external fabric, when you add the switch_user policy and provide the username and password, the password must be an encrypted string that is displayed in the show run command.
For example:
username admin password 5 $5$I4sapkBh$S7B7UcPH/iVTihLKH5sgldBeS3O2X1StQsvv3cmbYd1 role network-admin
In this case, the entered password should be 5$5$I4sapkBh$S7B7UcPH/iVTihLKH5sgldBeS3O2X1StQsvv3cmbYd1.
-
For the Cisco Network Insights for Resources (NIR) Release 2.1 and later, and flow telemetry, feature lldp command is one of the required configuration.
Cisco DCNM pushes feature lldp on the switches only for the Easy Fabric deployments, that is, for the eBGP routed fabric or VXLAN EVPN fabric.
Therefore, NIR users need to enable feature lldp on all the switches in the following scenarios:
-
External fabric in Monitored or Managed Mode
-
LAN Classic fabric in Monitored or Managed Mode (Applicable for DCNM 11.4(1) or later)
-
Creating External Fabric from Fabric Builder
Follow these steps to create an external fabric from Fabric Builder.
-
Click Control > Fabric Builder. The Fabric Builder page comes up.
-
Click the Create Fabric button. The Add Fabric screen comes up. The fields in this screen are:
Fabric Name - Enter the name of the external fabric.
Fabric Template - Choose External_Fabric.
When you choose the fabric template, the fabric creation screen for creating an external fabric comes up.
-
Fill up the General tab as shown below.
BGP AS # - Enter the BGP AS number.
Fabric Monitor Mode – Clear the check box if you want DCNM to manage the fabric. Keep the check box selected to enable a monitor only external fabric. DCNM supports only the monitor mode for Cisco IOS-XR family devices.
When you create an Inter-Fabric Connection from a VXLAN fabric to this external fabric, the BGP AS number is referenced as the external or neighbor fabric AS Number.
When an external fabric is set to Fabric Monitor Mode Only, you cannot deploy configurations on its switches. If you click Save & Deploy in the fabric topology screen, it displays an error message.
The configurations must be pushed for non-Nexus devices before you discover them in the fabric. You cannot push configurations in the monitor mode.
However, the following settings (available when you right-click the switch icon) are allowed:
-
Enter values in the fields under the Advanced tab.
vPC Peer Link VLAN - The vPC peer link VLAN ID is autopopulated. Update the field to reflect the correct value.
Power Supply Mode - Choose the appropriate power supply mode.
Enable MPLS Handoff: Select the check box to enable the MPLS Handoff feature. For more information, see the Border Provisioning Use Case in VXLAN BGP EVPN Fabrics - MPLS SR and LDP Handoff chapter.
Underlay MPLS Loopback Id: Specifies the underlay MPLS loopback ID. The default value is 101.
Enable AAA IP Authorization - Enables AAA IP authorization, when IP Authorization is enabled in the AAA Server
Enable DCNM as Trap Host - Select this check box to enable DCNM as a trap host.
Enable CDP for Bootstrapped Switch - Select the check box to enable CDP for bootstrapped switch.
Enable NX-API - Specifies enabling of NX-API on HTTPS. This check box is unchecked by default.
Enable NX-API on HTTP - Specifies enabling of NX-API on HTTP. This check box is unchecked by default. Enable this check box and the Enable NX-API check box to use HTTP. If you uncheck this check box, the applications that use NX-API and supported by Cisco DCNM, such as Endpoint Locator (EPL), Layer 4-Layer 7 services (L4-L7 services), VXLAN OAM, and so on, start using the HTTPS instead of HTTP.
Note
If you check the Enable NX-API check box and the Enable NX-API on HTTP check box, applications use HTTP.
Inband Mgmt: For External and Classic LAN Fabrics, this knob enables DCNM to import and manage of switches with inband connectivity (reachable over switch loopback, routed, or SVI interfaces) , in addition to management of switches with out-of-band connectivity (aka reachable over switch mgmt0 interface). The only requirement is that for Inband managed switches, there should be IP reachability from DCNM to the switches via the eth2 aka inband interface. For this purpose, static routes may be needed on the DCNM, that in turn can be configured via the Administration->Customization->Network Preferences option. After enabling Inband management, during discovery, provide the IPs of all the switches to be imported using Inband Management and set maximum hops to 0. DCNM has a pre-check that validates that the Inband managed switch IPs are reachable over the eth2 interface. Once the pre-check has passed, DCNM then discovers and learns about the interface on that switch that has the specified discovery IP in addition to the VRF that the interface belongs to. As part of the process of switch import/discovery, this information is captured in the baseline intent that is populated on the DCNM. For more information, see Inband Management in External Fabrics and LAN Classic Fabrics.
Note
Bootstrap or POAP is only supported for switches that are reachable over out-of-band connectivity, that is, over switch mgmt0. The various POAP services on the DCNM are typically bound to the eth1 or out-of-band interface. In scenarios, where DCNM eth0/eth1 interfaces reside in the same IP subnet, the POAP services are bound to both interfaces.
Enable Precision Time Protocol (PTP): Enables PTP across a fabric. When you select this check box, PTP is enabled globally and on core-facing interfaces. Additionally, the PTP Source Loopback Id and PTP Domain Id fields are editable. For more information, see Precision Time Protocol for External Fabrics and LAN Classic Fabrics.
PTP Source Loopback Id: Specifies the loopback interface ID Loopback that is used as the Source IP Address for all PTP packets. The valid values range from 0 to 1023. The PTP loopback ID cannot be the same as RP, Phantom RP, NVE, or MPLS loopback ID. Otherwise, an error will be generated. The PTP loopback ID can be the same as BGP loopback or user-defined loopback which is created from DCNM. If the PTP loopback ID is not found during Save & Deploy, the following error is generated:
Loopback interface to use for PTP source IP is not found. Please create PTP loopback interface on all the devices to enable PTP feature.
PTP Domain Id: Specifies the PTP domain ID on a single network. The valid values range from 0 to 127.
Fabric Freeform: You can apply configurations globally across all the devices discovered in the external fabric using this freeform field. The devices in the fabric should belong to the same device-type and the fabric should not be in monitor mode. The different device types are:
-
NX-OS
-
IOS-XE
-
IOS-XR
-
Others
Depending on the device types, enter the configurations accordingly. If some of the devices in the fabric do not support these global configurations, they will go out-of-sync or fail during the deployment. Hence, ensure that the configurations you apply are supported on all the devices in the fabric or remove the devices that do not support these configurations.
-
-
Fill up the Resources tab as shown below.
Subinterface Dot1q Range - The subinterface 802.1Q range and the underlay routing loopback IP address range are autopopulated.
Underlay Routing Loopback IP Range - Specifies loopback IP addresses for the protocol peering.
Underlay MPLS Loopback IP Range: Specifies the underlay MPLS SR or LDP loopback IP address range.
The IP range should be unique, that is, it should not overlap with IP ranges of the other fabrics.
Enable AAA IP Authorization - Enables AAA IP authorization, when IP Authorization is enabled in the AAA Server
Enable DCNM as Trap Host - Select this check box to enable DCNM as a trap host.
-
Fill up the Configuration Backup tab as shown below.
The fields on this tab are:
Hourly Fabric Backup: Select the check box to enable an hourly backup of fabric configurations and the intent.
You can enable an hourly backup for fresh fabric configurations and the intent as well. If there is a configuration push in the previous hour, DCNM takes a backup. In case of the external fabric, the entire configuration on the switch is not converted to intent on DCNM as compared to the VXLAN fabric. Therefore, for the external fabric, both intent and running configuration are backed up.
Intent refers to configurations that are saved in DCNM but yet to be provisioned on the switches.
The hourly backups are triggered during the first 10 minutes of the hour.
Scheduled Fabric Backup: Check the check box to enable a daily backup. This backup tracks changes in running configurations on the fabric devices that are not tracked by configuration compliance.
Scheduled Time: Specify the scheduled backup time in a 24-hour format. This field is enabled if you check the Scheduled Fabric Backup check box.
Select both the check boxes to enable both back up processes.
The backup process is initiated after you click Save.
The scheduled backups are triggered exactly at the time you specify with a delay of up to two minutes. The scheduled backups are triggered regardless of the configuration deployment status.
You can also initiate the fabric backup in the fabric topology window. Click Backup Now in the Actions pane.
Pointers for hourly and scheduled backup:
-
The backups contain running configuration and intent pushed by DCNM. Configuration compliance forces the running config to be the same as the DCNM config. Note that for the external fabric, only some configurations are part of intent and the remaining configurations are not tracked by DCNM. Therefore, as part of backup, both DCNM intent and running config from switch are captured.
-
-
Click the Bootstrap tab.
Enable Bootstrap - Select this check box to enable the bootstrap feature. After you enable bootstrap, you can enable the DHCP server for automatic IP address assignment using one of the following methods:
-
External DHCP Server: Enter information about the external DHCP server in the Switch Mgmt Default Gateway and Switch Mgmt IP Subnet Prefix fields.
-
Local DHCP Server: Enable the Local DHCP Server check box and enter details for the remaining mandatory fields.
Enable Local DHCP Server - Select this check box to initiate enabling of automatic IP address assignment through the local DHCP server. When you select this check box, all the remaining fields become editable.
DHCP Version – Select DHCPv4 or DHCPv6 from this drop-down list. When you select DHCPv4, the Switch Mgmt IPv6 Subnet Prefix field is disabled. If you select DHCPv6, the Switch Mgmt IP Subnet Prefix is disabled.
Note
Cisco DCNM IPv6 POAP is not supported with Cisco Nexus 7000 Series Switches. Cisco Nexus 9000 and 3000 Series Switches support IPv6 POAP only when switches are either L2 adjacent (eth1 or out-of-band subnet must be a /64) or they are L3 adjacent residing in some IPv6 /64 subnet. Subnet prefixes other than /64 are not supported.
If you do not select this check box, DCNM uses the remote or external DHCP server for automatic IP address assignment.
DHCP Scope Start Address and DHCP Scope End Address - Specifies the first and last IP addresses of the IP address range to be used for the switch out of band POAP.
Switch Mgmt Default Gateway - Specifies the default gateway for the management VRF on the switch.
Switch Mgmt IP Subnet Prefix - Specifies the prefix for the Mgmt0 interface on the switch. The prefix should be between 8 and 30.
DHCP scope and management default gateway IP address specification - If you specify the management default gateway IP address 10.0.1.1 and subnet mask 24, ensure that the DHCP scope is within the specified subnet, between 10.0.1.2 and 10.0.1.254.
Switch Mgmt IPv6 Subnet Prefix - Specifies the IPv6 prefix for the Mgmt0 interface on the switch. The prefix should be between 112 and 126. This field is editable if you enable IPv6 for DHCP.
Enable AAA Config - Select this check box to include AAA configs from Advanced tab during device bootup.
Bootstrap Freeform Config - (Optional) Enter other commands as needed. For example, if you are using AAA or remote authentication-related configurations, add these configurations in this field to save the intent. After the devices boot up, they contain the intent defined in the Bootstrap Freeform Config field.
Copy-paste the running-config to a freeform config field with correct indentation, as seen in the running configuration on the NX-OS switches. The freeform config must match the running config. For more information, see Resolving Freeform Config Errors in Switches.
DHCPv4/DHCPv6 Multi Subnet Scope - Specifies the field to enter one subnet scope per line. This field is editable after you check the Enable Local DHCP Server check box.
The format of the scope should be defined as:
DHCP Scope Start Address, DHCP Scope End Address, Switch Management Default Gateway, Switch Management Subnet Prefix
For example: 10.6.0.2, 10.6.0.9, 10.6.0.1, 24
-
-
Click ThousandEyes Agent tab. This feature is supported on Cisco DCNM Release 11.5(3) only. For more information, refer to Configuring Global Settings for ThousandEyes Enterprise Agent.
The fields on this tab are:
Note
The fabric settings for ThousandEyes Agent overwrites the global settings and applies the same configuration for all the ThousandEyes Agent installed on switches in that fabric.
-
Enable Fabric Override for ThousandEyes Agent Installation: Select the check box to enable the ThousandEyes Enterprise Agent on the fabric.
-
ThousandEyes Account Group Token: Specifies ThousandEyes Enterprise Agent account group token for installation.
-
VRF on Switch for ThousandEyes Agent Collector Reachability: Specifies the VRF data which provides internet reachability.
-
DNS Domain: Specifies the switch DNS domain configuration.
-
DNS Server IPs: Specifies the comma separated list of IP addresses (v4/v6) of Domain Name System (DNS) server. You can enter a maximum of three IP addresses for the DNS Server.
-
NTP Server IPs: Specifies comma separated list of IP addresses (v4/v6) of Network Time Protocol (NTP) server. You can enter a maximum of three IP addresses for the NTP Server.
-
Enable Proxy for Internet Access: Select the check box to enable the proxy setting for NX-OS switch internet access.
-
Proxy Information: Specifies the proxy server port information.
-
Proxy Bypass: Specifies the server list for which proxy is bypassed.
-
-
Click Save.
After the external fabric is created, the external fabric topology page comes up.
After creating the external fabric, add switches to it.
Add Switches to the External Fabric
-
Click Add switches. The Inventory Management screen comes up.
You can also add switches by clicking Tabular View > Switches > + .
-
Enter the IP address (Seed IP) of the switch.
-
Choose the device type from the Device Type drop-down list.
The options are NX-OS, IOS XE, IOS XR, and Other.
-
Choose NX-OS to discover a Cisco Nexus switch.
-
Choose IOS XE to discover a CSR device.
-
Choose IOS XR to discover an ASR device.
-
Choose Other to discover non-Cisco devices.
Click the appropriate radio button. Refer the Connecting Cisco Data Center and a Public Cloud chapter for more information on adding Cisco CSR 1000v.
Refer the Adding non-Nexus Devices to External Fabrics section for more information on adding other non-Nexus devices.
Config compliance is disabled for all non-Nexus devices except for Cisco CSR 1000v.
-
-
Enter the administrator username and password of the switch.
-
Click Start discovery at the bottom part of the screen. The Scan Details section comes up shortly. Since the Max Hops field was populated with 2, the switch with the specified IP address and switches two hops from it are populated.
-
Select the check boxes next to the concerned switches and click Import into fabric.
You can discover multiple switches at the same time. The switches must be properly cabled and connected to the DCNM server and the switch status must be manageable.
The switch discovery process is initiated. The Progress column displays the progress. After DCNM discovers the switch, the screen closes and the fabric screen comes up again. The switch icons are seen at the centre of the fabric screen.
-
Click Refresh topology to view the latest topology view.
-
External Fabric Switch Settings - The settings for external fabric switches vary from the VXLAN fabric switch settings. Right-click on the switch icon and set or update switch options.
The options are:
Set Role – By default, no role is assigned to an external fabric switch. The allowed roles are Edge Router and Core Router. Assign the Core Router role for a Multi-Site Inter-Fabric Connection (IFC) and the Edge Router role for a VRF Lite IFC between the external fabric and VXLAN fabric border devices.
Note
Changing of switch role is allowed only before executing Save & Deploy.
Modes – Active/Operational mode.
vPC Pairing – Select a switch for vPC and then select its peer.
Manage Interfaces – Deploy configurations on the switch interfaces.
Straight-through FEX, Active/Active FEX, and breakout of interfaces are not supported for external fabric switch interfaces.
View/edit Policies – Add, update, and delete policies on the switch. The policies you add to a switch are template instances of the templates available in the template library. After creating policies, deploy them on the switch using the Deploy option available in the View/edit Policies screen.
History – View per switch deployment history.
Preview Config - View the pending configuration and the side-by-side comparison of the running and expected configuration.
Deploy Config – Deploy per switch configurations.
Discovery - You can use this option to update the credentials of the switch, reload the switch, rediscover the switch, and remove the switch from the fabric.
-
Click Save & Deploy at the top right part of the screen. The template and interface configurations form the configuration provisioning on the switches.
When you click Save & Deploy, the Configuration Deployment screen comes up.
-
Click Deploy Config at the bottom part of the screen to initiate pending configuration onto the switch.
-
Close the screen after deployment is complete.
Note
If a switch in an external fabric does not accept default credentials, you should perform one of the following actions:
-
Remove the switch in the external fabric from inventory, and then rediscover.
-
LAN discovery uses both SNMP and SSH, so both passwords need to be the same. You need to change the SSH password to match the SNMP password on the switch. If SNMP authentication fails, discovery is stopped with authentication error. If SNMP authentication passes but SSH authentication fails, DCNM discovery continues, but the switch status shows a warning for the SSH error.
-
Move an External Fabric Under an MSD Fabric
You should go to the MSD fabric page to associate an external fabric as its member.
-
Click Control > Fabric Builder to go to the Fabric Builder screen.
-
Click within the MSD-Parent-Fabric box to go to its topology screen.
-
In the topology screen, go to the Actions panel and click Move Fabrics.
The Move Fabric screen comes up. It contains a list of fabrics. The external fabric is displayed as a standalone fabric.
-
Select the radio button next to the external fabric and click Add.
Now, in the Scope drop-down box at the top right, you can see that the external fabric appears under the MSD fabric.
-
Click ← at the top left part of the screen to go to the Fabric Builder screen. In the MSD fabric box’s Member Fabrics field, the external fabric is displayed.
External Fabric Depiction in an MSD Fabric Topology
The MSD topology screen displays MSD member fabrics and external fabrics together. The external fabric External65000 is displayed as part of the MSD topology.
Note |
When you deploy networks or VRFs for the VXLAN fabric, the deployment page (MSD topology view) shows the VXLAN and external fabrics that are connected to each other. |
External Fabric Switch Operations
In the external fabric topology screen, click Tabular view option in the Actions panel, at the left part of the screen. The Switches | Links screen comes up.
The Switches tab is for managing switch operations and the Links tab is for viewing fabric links. Each row represents a switch in the external fabric, and displays switch details, including its serial number.
The buttons at the top of the table are explained, from left to right direction. Some options are also available when you right-click the switch icon. However, the Switches tab enables you to provision configurations on multiple switches (for adding and deploying policies, and so on) simultaneously.
-
Add switches to the fabric. This option is also available in the topology page (Add switches option in Actions panel).
-
Initiate the switch discovery process by DCNM afresh.
-
Update device credentials such as authentication protocol, username, and password.
-
Reload the switch.
-
Remove the switch from the fabric.
-
View/edit Policies – Add, update, and delete a policy on multiple switches simultaneously. The policies are template instances of templates in the template library. After creating a policy, deploy it on the switches using the Deploy option available in the View/edit Policies screen.
Note
If you select multiple switches and deploy a policy instance, then it will be deployed on all the selected switches.
-
Manage Interfaces – Deploy configurations on the switch interfaces.
-
History – View deployment history on the selected switch.
-
Deploy – Deploy switch configurations.
External Fabric Links
You can only view and delete external fabric links. You cannot create links or edit them.
To delete a link in the external fabric, do the following:
-
Go to the topology screen and click the Tabular view option in the Actions panel, at the left part of the screen.
The Switches | Links screen comes up.
-
Choose one or more check boxes and click the Delete icon at the top left.
The links are deleted.
Move Neighbor Switch to External Fabric
-
Click Add switches. The Inventory Management screen comes up.
-
Click Move Neighbor Switches tab.
-
Select the switch and click Move Neighbor.
To delete a neighbor, select a switch and click Delete Neighbor.
Discovering New Switches
Procedure
Step 1 |
Power on the new switch in the external fabric after ensuring that it is cabled to the DCNM server. Boot the Cisco NX-OS and setup switch credentials. |
||
Step 2 |
Execute the write, erase, and reload commands on the switch. Choose Yes to both the CLI commands that prompt you to choose Yes or No. |
||
Step 3 |
On the DCNM UI, choose Control > Fabric Builder. The Fabric Builder screen is displayed. It contains a list of fabrics wherein a rectangular box represents each fabric. |
||
Step 4 |
Click Edit Fabric icon at the top right part of the fabric box. The Edit Fabric screen is displayed. |
||
Step 5 |
Click the Bootstrap tab and update the DHCP information. |
||
Step 6 |
Click Save at the bottom right part of the Edit Fabric screen to save the settings. |
||
Step 7 |
In the Fabric Builder screen, click within the fabric box. The fabric topology screen appears. |
||
Step 8 |
In the fabric topology screen, from the Actions panel at the left part of the screen, click Add switches. The Inventory Management screen comes up. |
||
Step 9 |
Click the POAP tab. In an earlier step, the reload command was executed on the switch. When the switch restarts to reboot, DCNM retrieves the serial number, model number, and version from the switch and displays them on the Inventory Management along screen. Also, an option to add the management IP address, hostname, and password are made available. If the switch information is not retrieved, refresh the screen using the Refresh icon at the top right part of the screen.
Select the checkbox next to the switch and add switch credentials: IP address and host name. Based on the IP address of your device, you can either add the IPv4 or IPv6 address in the IP Address field. Beginning with Release 11.2(1), you can provision devices in advance. To pre-provision devices, refer to Pre-provisioning a Device. |
||
Step 10 |
In the Admin Password and Confirm Admin Password fields, enter and confirm the admin password. This admin password is applicable for all the switches displayed in the POAP window.
|
||
Step 11 |
(Optional) Use discovery credentials for discovering switches. |
||
Step 12 |
Click Bootstrap at the top right part of the screen. DCNM provisions the management IP address and other credentials to the switch. In this simplified POAP process, all ports are opened up. |
||
Step 13 |
After the bootstrapping is complete, close the Inventory Management screen to go to the fabric topology screen. |
||
Step 14 |
In the fabric topology screen, from the Actions panel at the left part of the screen, click Refresh Topology. After the added switch completes POAP, the fabric builder topology screen displays the added switch with some physical connections. |
||
Step 15 |
Monitor and check the switch for POAP completion. |
||
Step 16 |
Click Save & Deploy at the top right part of the fabric builder topology screen to deploy pending configurations (such as template and interface configurations) onto the switches.
During fabric creation, if you have entered AAA server information (in the Manageability tab), you must update the AAA server password on each switch. Else, switch discovery fails. |
||
Step 17 |
After the pending configurations are deployed, the Progress column displays 100% for all switches. |
||
Step 18 |
Click Close to return to the fabric builder topology. |
||
Step 19 |
Click Refresh Topology to view the update. All switches must be in green color indicating that they are functional. The switch and the link are discovered in DCNM. Configurations are built based on various policies (such as fabric, topology, and switch generated policies). The switch image (and other required) configurations are enabled on the switch. |
||
Step 20 |
Right-click and select History to view the deployed configurations. Click the Success link in the Status column for more details. An example: |
||
Step 21 |
On the DCNM UI, the discovered switches can be seen in the fabric topology. Up to this step, the POAP is completed with basic settings. All the interfaces are set to trunk ports. You must setup interfaces through the Control > Interfaces option for any additional configurations, but not limited to the following:
|
Adding non-Nexus Devices to External Fabrics
You can discover non-Nexus devices in an external fabric. Refer the Cisco DCNM Compatibility Matrix to see the non-Nexus devices supported by Cisco DCNM.
Only Cisco Nexus switches support SNMP discovery by default. Hence, configure all the non-Nexus devices before adding it to the external fabric. Configuring the non-Nexus devices includes configuring SNMP views, groups, and users. See the Configuring non-Nexus Devices for Discovery section for more information.
Cisco CSR 1000v is discovered using SSH. Cisco CSR 1000v does not need SNMP support because it can be installed in clouds where SNMP is blocked for security reasons. See the Connecting Cisco Data Center and a Public Cloud chapter to see a use case to add Cisco CSR 1000v, Cisco IOS XE Gibraltar 16.10.x to an external fabric.
However, Cisco DCNM can only access the basic device information like system name, serial number, model, version, interfaces, up time, and so on. Cisco DCNM does not discover non-Nexus devices if the hosts are part of CDP or LLDP.
The settings that are not applicable for non-Nexus devices appear blank, even if you get many options when you right-click a non-Nexus device in the fabric topology window. You cannot add or edit interfaces for ASR 9000 Series Routers and Arista switches.
From Cisco DCNM, Release 11.4(1), you can add IOS-XE devices like Cisco Catalyst 9000 Series switches and Cisco ASR 1000 Series Routers as well to external fabrics.
Configuring non-Nexus Devices for Discovery
Before discovering any non-Nexus device in Cisco DCNM, configure it on the switch console.
Configuring IOS-XE Devices for Discovery
Before you discover the Cisco IOS-XE devices in DCNM, perform the following steps:
Procedure
Step 1 |
Run the following SSH commands on the switch console.
|
Step 2 |
Run the following command in DCNM console to perform an SNMP walk.
|
Step 3 |
Run the following SNMP command on the switch console.
|
Configuring Arista Devices for Discovery
Enable Privilege Exec mode using the following command:
switch> enable
switch#
switch# show running confiruation | grep aaa /* to view the authorization*/
aaa authorization exec default local
switch# configure terminal
switch (config)# username dcnm privilege 15 role network-admin secret cisco123
snmp-server view view_name SNMPv2 included
snmp-server view view_name SNMPv3 included
snmp-server view view_name default included
snmp-server view view_name entity included
snmp-server view view_name if included
snmp-server view view_name iso included
snmp-server view view_name lldp included
snmp-server view view_name system included
snmp-server view sys-view default included
snmp-server view sys-view ifmib included
snmp-server view sys-view system included
snmp-server community private ro
snmp-server community public ro
snmp-server group group_name v3 auth read view_name
snmp-server user username group_name v3 auth md5 password priv aes password
Note |
SNMP password should be same as the password for username. |
You can verify the configuration by running the show run command, and view the SNMP view output by running the show snmp view command.
Show Run Command
switch (config)# snmp-server engineID local f5717f444ca824448b00
snmp-server view view_name SNMPv2 included
snmp-server view view_name SNMPv3 included
snmp-server view view_name default included
snmp-server view view_name entity included
snmp-server view view_name if included
snmp-server view view_name iso included
snmp-server view view_name lldp included
snmp-server view view_name system included
snmp-server view sys-view default included
snmp-server view sys-view ifmib included
snmp-server view sys-view system included
snmp-server community private ro
snmp-server community public ro
snmp-server group group_name v3 auth read view_name
snmp-server user user_name group_name v3 localized f5717f444ca824448b00 auth md5 be2eca3fc858b62b2128a963a2b49373 priv aes be2eca3fc858b62b2128a963a2b49373
!
spanning-tree mode mstp
!
service unsupported-transceiver labs f5047577
!
aaa authorization exec default local
!
no aaa root
!
username admin role network-admin secret sha512 $6$5ZKs/7.k2UxrWDg0$FOkdVQsBTnOquW/9AYx36YUBSPNLFdeuPIse9XgyHSdEOYXtPyT/0sMUYYdkMffuIjgn/d9rx/Do71XSbygSn/
username cvpadmin role network-admin secret sha512 $6$fLGFj/PUcuJT436i$Sj5G5c4y9cYjI/BZswjjmZW0J4npGrGqIyG3ZFk/ULza47Kz.d31q13jXA7iHM677gwqQbFSH2/3oQEaHRq08.
username dcnm privilege 15 role network-admin secret sha512 $6$M48PNrCdg2EITEdG$iiB880nvFQQlrWoZwOMzdt5EfkuCIraNqtEMRS0TJUhNKCQnJN.VDLFsLAmP7kQBo.C3ct4/.n.2eRlcP6hij/
Show SNMP View Command
configure terminal# show snmp view
view_name SNMPv2 - included
view_name SNMPv3 - included
view_name default - included
view_name entity - included
view_name if - included
view_name iso - included
view_name lldp - included
view_name system - included
sys-view default - included
sys-view ifmib - included
sys-view system - included
leaf3-7050sx#show snmp user
User name : user_name
Security model : v3
Engine ID : f5717f444ca824448b00
Authentication : MD5
Privacy : AES-128
Group : group_name
Configuring Cisco IOS-XR Devices for Discovery
Run the following commands in the switch console to configure IOS-XR devices:
switch# configure terminal
switch (config)# snmp-server view view_name cisco included
snmp-server view view_name mib-2 included
snmp-server group group_name v3 auth read view_name write view_name
snmp-server user user_name group_name v3 auth md5 password priv des56 password SystemOwner
Note |
SNMP password should be same as password for username. |
You can verify the configuration by running the show run command.
Configuration and Verification of Cisco IOS-XR Devices
RP/0/RSP0/CPU0:ios(config)#snmp-server view view_name cisco included
RP/0/RSP0/CPU0:ios(config)#snmp-server view view_name mib-2 included
RP/0/RSP0/CPU0:ios(config)#snmp-server group group_name v3 auth read view_name write view_name
RP/0/RSP0/CPU0:ios(config)#snmp-server user user_name group_name v3 auth md5 password priv des56 password SystemOwner
RP/0/RSP0/CPU0:ios(config)#commit Day MMM DD HH:MM:SS Timezone
RP/0/RSP0/CPU0:ios(config)#
RP/0/RSP0/CPU0:ios(config)#show run snmp-server Day MMM DD HH:MM:SS Timezone snmp-server user user_name group1 v3 auth md5 encrypted 10400B0F3A4640585851 priv des56 encrypted 000A11103B0A59555B74 SystemOwner
snmp-server view view_name cisco included
snmp-server view view_name mib-2 included
snmp-server group group_name v3 auth read view_name write view_name
Discovering non-Nexus Devices in an External Fabric
To add non-Nexus devices to an external fabric in the fabric topology window, perform the following steps:
Before you begin
Ensure that the configurations are pushed for non-Nexus devices before adding them to an external fabric. You cannot push configurations in a fabric in the monitor mode.
Procedure
Step 1 |
Click Add switches in the Actions pane. The Inventory Management dialog box appears. |
||||||||||||
Step 2 |
Enter values for the following fields under the Discover Existing Switches tab:
|
||||||||||||
Step 3 |
Click Start Discovery. The Scan Details section appears with the switch details populated. |
||||||||||||
Step 4 |
Check the check boxes next to the switches you want to import. |
||||||||||||
Step 5 |
Click Import into fabric. The switch discovery process is initiated. The Progress column displays the progress. Discovering devices takes some time. A pop-up message appears at the bottom-right about the device discovery after the discovery progress is 100%, or done. For example: <ip-address> added for discovery. |
||||||||||||
Step 6 |
Click Close. The fabric topology window appears with the switches. |
||||||||||||
Step 7 |
(Optional) Click Refresh topology to view the latest topology view. |
||||||||||||
Step 8 |
(Optional) Click Tabular view in the Actions pane. The switches and links window appears, where you can view the scan details. The discovery status is discovering in red with a warning icon next to it if the discovery is in progress. |
||||||||||||
Step 9 |
(Optional) View the details of the device. After the discovery of the device:
|
What to do next
Pre-provisioning a Device
Note |
Ensure that you enter DHCP details in the Bootstrap tab in the fabric settings. |
-
The pre-provisioned devices support the following configurations in DCNM:
-
Base management
-
vPC Pairing
-
Intra-Fabric links
-
Ethernet ports
-
Port-channel
-
vPC
-
ST FEX
-
AA FEX
-
Loopback
-
Overlay network configurations
-
-
The pre-provisioned devices do not support the following configurations in DCNM:
-
Inter-Fabric links
-
Sub-interface
-
Interface breakout configuration
-
-
When a device is being pre-provisioned has breakout links, you need to specify the corresponding breakout command along with the switch's model and gateway in the Data field in the Add a new device to pre-provisioning window in order to generate the breakout PTI.
Note the following guidelines:
-
Multiple breakout commands can be separated by a semicolon (;).
-
The definitions of the fields in the data JSON object are as follows:
-
modulesModel: (Mandatory) Specifies the switch module’s model information.
-
gateway: (Mandatory) Specifies the default gateway for the management VRF on the switch. This field is required to create the intent to pre-provision devices. You must enter the gateway even if it is in the same subnet as DCNM to create the intent as part of pre-provisioning a device.
-
breakout: (Optional) Specifies the breakout command provided in the switch.
-
portMode: (Optional) Specifies the port mode of the breakout interface.
-
The examples of the values in the Data field are as follows:
-
{"modulesModel": ["N9K-C93180LC-EX"], "gateway": "10.1.1.1/24"}
-
{"modulesModel": ["N9K-C93180LC-EX"],"breakout": "interface breakout module 1 port 1 map 10g-4x", "portMode": "hardware profile portmode 4x100G+28x40G", "gateway": "172.22.31.1/24" }
-
{"modulesModel": ["N9K-X9736C-EX", "N9K-X9732C-FX", "N9K-C9516-FM-E2", "N9K-C9516-FM-E2", "N9K-C9516-FM-E2", "N9K-C9516-FM-E2", "N9K-SUP-B+", "N9K-SC-A", "N9K-SC-A"], "gateway": "172.22.31.1/24"}
-
{"breakout":"interface breakout module 1 port 50 map 10g-4x" , "gateway": "172.16.1.1/24", "modulesModel": ["N9K-C93180YC-EX "]}
-
{"modulesModel": ["N9K-X9732C-EX", "N9K-X9732C-EX", "N9K-C9504-FM-E", "N9K-C9504-FM-E", "N9K-SUP-B", "N9K-SC-A", "N9K-SC-A"], "gateway": "172.29.171.1/24", "breakout":"interface breakout module 1 port 1,11,19 map 10g-4x; interface breakout module 1 port 7 map 25g-4x"}
-
{"modulesModel": ["N9K-C93180LC-EX"], "gateway": "10.1.1.1/24","breakout": "interface breakout module 1 port 1-4 map 10g-4x","portMode": "hardware profile portmode 48x25G + 2x100G + 4x40G"}
-
Procedure
Step 1 |
Click Control > Fabric Builder. The Fabric Builder screen is displayed. |
||
Step 2 |
Click within the fabric box. |
||
Step 3 |
From the Actions panel, click the Add switches option. The Inventory Management screen is displayed. |
||
Step 4 |
Click the POAP tab. |
||
Step 5 |
In the POAP tab, do the following:
IP Address: Specify the IPv4 or IPv6 address of the new device. Serial Number: The serial number for the new device. Serial number is found in the Cisco Build of Material Purchase and you can refer to these values while using the pre-provisioning feature. For information about the Data field, see the examples provided in guidelines. The device details appear in the POAP screen. You can add more devices for pre-provisioning. At the top left part of the window, Export and Import icons are provided to export and import the .csv file that contains the switch information. Using the Import option, you can pre-provision multiple devices. Add new devices’ information in the .csv file with all the mandatory fields (SerialNumber, Model, version, IpAddress, Hostname, and Data fields [JSON Object]). The Data column consists of the model name of the module to identify the hardware type from the fabric template. A .csv file screenshot: |
||
Step 6 |
Enter the administration password in the Admin Password and Confirm Admin Password fields. |
||
Step 7 |
Select the device(s) and click Bootstrap at the top right part of the screen. The leaf1 device appears in the fabric topology. From the Actions panel, click Tabular View. You cannot deploy the fabric till the status of all the pre-provisioned switch(es) are displayed as ok under the Discovery Status column.
When you connect leaf1 to the fabric, the switch is provisioned with the IP address 10.1.1.1. |
||
Step 8 |
Navigate to Fabric Builder and set roles for the device. Create intra-link policy using one of the templates:
Click Save & Deploy. Configuration for the switches are captured in corresponding PTIs and can be seen in the View/Edit Policies window. |
||
Step 9 |
To bring in the physical device, you can follow the manual RMA or POAP RMA procedure. For more information, see Return Material Authorization (RMA). If you use the POAP RMA procedure, ignore the error message of failing to put the device into maintenance mode due to no connectivity since it is expected to have no connectivity to a non-existing device. You need to click Save & Deploy in the fabric after one or more switches are online to provision the host ports. This action must be performed before overlays are provisioned for the host port attachment. |
Pre-provisioning an Ethernet Interface
From DCNM Release 11.4(1), you can pre-provision Ethernet interfaces in the Interface window. This pre-provisioning feature is supported in the Easy, External, and eBGP fabrics. You can add Ethernet interfaces to only pre-provisioned devices before they are discovered in DCNM.
Note |
Before attaching a network/VRF, you must pre-provision the Ethernet interface before adding it to Port-channels, vPCs, ST FEX, AA FEX, loopback, subinterface, tunnel, ethernet, and SVI configurations. |
Before you begin
Procedure
Step 1 |
Navigate to the fabric containing the pre-provisioned device from the Fabric Builder window. |
||
Step 2 |
Right click the pre-provisioned device and select Manage Interfaces. You can also navigate to the Interfaces window by selecting Control > Fabrics > Interfaces. From the Scope drop-down list, select the fabric containing the pre-provisioned device. |
||
Step 3 |
Click Add. |
||
Step 4 |
Enter all the required details in the Add Interface window. Type: Select Ethernet from this drop-down list. Select a device: Select the pre-provisioned device.
Enter Interface Name: Enter a valid interface name based on the module type. For example, Ethernet1/1, eth1/1, or e1/1. The interface with same name should be available on the device after it is added. Policy: Select a policy that should be applied on the interface. For more information, see Adding Interfaces. |
||
Step 5 |
Click Save. |
||
Step 6 |
Click Preview to check the expected configuration that will be deployed to the switch after it is added.
|
Creating a vPC Setup
Procedure
Step 1 |
Right-click one of the two designated vPC switches and choose vPC Pairing. The Select vPC peer dialog box comes up. It contains a list of potential peer switches. Ensure that the Recommended column for the vPC peer switch is updated as true.
|
||
Step 2 |
Click the radio button next to the vPC peer switch and choose vpc_pair from the vPC Pair Template drop-down list. Only templates with the VPC_PAIR template sub type are listed here. The vPC Domain and vPC Peerlink tabs appear. You must fill up the fields in the tabs to create the vPC setup. The description for each field is displayed at the extreme right. vPC Domain tab: Enter the vPC domain details. vPC+: If the switch is part of a FabricPath vPC + setup, enable this check box and enter the FabricPath switch ID field. Configure VTEPs: Check this check box to enter the source loopback IP addresses for the two vPC peer VTEPs and the loopback interface secondary IP address for NVE configuration. NVE interface: Enter the NVE interface. vPC pairing will configure only the source loopback interface. Use the freeform interface manager for additional configuration. NVE loopback configuration: Enter the IP address with the mask. vPC pairing will only configure primary and secondary IP address for loopback interface. Use the freeform interface manager for additional configuration. vPC Peerlink tab: Enter the vPC peer-link details. Switch Port Mode: Choose trunk or access or fabricpath. If you select trunk, then corresponding fields (Trunk Allowed VLANs and Native VLAN) are enabled. If you select access, then the Access VLAN field is enabled. If you select fabricpath, then the trunk and access port related fields are disabled. |
||
Step 3 |
Click Save. The fabric topology window appears. The vPC setup is created. To update vPC setup details, do the following:
|
Undeploying a vPC Setup
Procedure
Step 1 |
Right-click a vPC switch and choose vPC Pairing. The vPC peer screen comes up. |
||
Step 2 |
Click Unpair at the bottom right part of the screen. The vPC pair is deleted and the fabric topology window appears. |
||
Step 3 |
Click Save & Deploy. The Config Deployment dialog box appears. |
||
Step 4 |
(Optional) Click the value under the Preview Config column. View the pending configuration in the Config Preview dialog box. The following configuration details are deleted on the switch when you unpair: vPC feature, vPC domain, vPC peerlink, vPC peerlink member ports, loopback secondary IPs, and host vPCs. However, the host vPCs and port channels are not removed. Delete these port channels from the Interfaces window if required.
|
Multi-Site Domain for VXLAN BGP EVPN Fabrics
A Multi-Site Domain (MSD) is a multifabric container that is created to manage multiple member fabrics. An MSD is a single point of control for definition of overlay networks and VRFs that are shared across member fabrics. When you move fabrics (that are designated to be part of the multifabric overlay network domain) under the MSD as member fabrics, the member fabrics share the networks and VRFs created at the MSD-level. This way, you can consistently provision network and VRFs for different fabrics, at one go. It significantly reduces the time and complexity involving multiple fabric provisionings.
Since server networks and VRFs are shared across the member fabrics (as one stretched network), the new networks and VRFs provisioning function is provided at the MSD fabric level. Any new network and VRF creation is only allowed for the MSD. All member fabrics inherit any new network and VRF created for the MSD.
In DCNM 11.1(1) release, in addition to member fabrics, the topology view for the MSD fabric is introduced. This view displays all member fabrics, and how they are connected to each other, in one view.
Also, a deployment view is introduced for the MSD fabric. You can deploy overlay networks (and VRFs) on member fabrics from a single topology deployment screen, instead of visiting each member fabric deployment screen separately and deploying.
Note |
|
A few fabric-specific terms:
-
Standalone fabric: A fabric that is not part of an MSD is referred as a standalone fabric from the MSD perspective. Before the MSD concept, all fabrics were considered standalone, though two or more such fabrics can be connected with each other.
-
Member fabrics: Fabrics that are part of an MSD are called member fabrics or members. Create a standalone fabric (of the type Easy_Fabric) first and then move it within an MSD as a member fabric.
When a standalone fabric is added to the MSD, the following actions take place:
-
The standalone fabric's relevant attributes and the network and VRF definitions are checked against that of the MSD. If there is a conflict, then the standalone fabric addition to the MSD fails. If there are no conflicts, then the standalone fabric becomes a member fabric for the MSD. If there is a conflict, the exact conflicts are logged in the pending errors log for the MSD fabric. You can remedy the conflicts and then attempt to add the standalone fabric to the MSD again.
-
All the VRFs and networks definitions from the standalone fabric that do not have presence in the MSD are copied over to the MSD and in turn inherited to each of its other existing member fabrics.
-
The VRFs (and their definitions) from the MSD (such as the MSD's VRF, and L2 and L3 VNI parameters that do not have presence in the standalone fabric) are inherited into the standalone fabric that just became a member.
Fabric and Switch Instance Variables
While the MSD provisions a global range of network and VRF values, some parameters are fabric-specific and some parameters are switch-specific. The parameters are called fabric instance and switch instance variables.
Fabric instance values can only be edited or updated in the fabric context from the VRFs and Networks window. The appropriate fabric should be selected in the SCOPE drop-down list to edit the fabric instance values. Some of the examples of fabric instance variables are BGP ASN, Multicast group per network or VRF, etc. For information about editing multicast group address, see Editing Networks in the Member Fabric.
Switch instance values can be edited on deployment of the network on the switch. For example, VLAN ID.
MSD and Member Fabric Process Flow
An MSD has multiple sites (and hence, multiple member fabrics under an MSD). VRFs and networks are created for the MSD and get inherited by the member fabrics. For example, VRF-50000 (and L3 network with ID 50000), and L2 networks with IDs 30000 and 30001 are created for the MSD, in one go.
A high-level flow chart of the MSD and member fabric creation and MSD-to-member fabric inheritance process:
The sample flow explained the inheritance from the MSD to one member. An MSD has multiple sites (and hence, multiple member fabrics under an MSD). A sample flow from an MSD to multiple members:
In this example, VRF-50000 (and L3 network with ID 50000), and L2 networks with IDs 30000 and 30001 are created in one go. Networks and VRFs are deployed on the member fabric switches, one after another, as depicted in the image.
In DCNM 11.1(1), you can provision overlay networks through a single MSD deployment screen.
Note |
If you move a standalone fabric with existing networks and VRFs to an MSD, DCNM does appropriate validation. This is explained in detail in an upcoming section. |
Upcoming sections in the document explain the following:
-
Creation of an MSD fabric.
-
Creation of a standalone fabric (as a potential member) and its movement under the MSD as a member.
-
Creation of networks and VRFs in the MSD and their inheritance to the member fabrics.
-
Deployment of networks and VRFs from the MSD and member fabric topology views.
-
Other scenarios for fabric movement:
-
Standalone fabric with existing networks and VRFs to an MSD fabric.
-
Member fabric from one MSD to another.
-
Creating an MSD Fabric and Associating Member Fabrics to It
The process is explained in two steps:
-
Create an MSD fabric.
-
Create a new standalone fabric and move it under the MSD fabric as a member fabric.
Creating an MSD Fabric
-
Click Control > Fabric Builder.
The Fabric Builder screen comes up. When you view the screen for the first time, the Fabrics section has no entries. After you create a fabric, it is displayed on the Fabric Builder screen, wherein a rectangular box represents each fabric.
A standalone or member fabric contains Switch_Fabric in the Type field, its AS number in the ASN field and mode of replication, Multicast or Ingress Replication, in the Replication Mode field. Since no device or network traffic is associated with an MSD fabric as it is a container, it does not have these fields.
-
Click the Create Fabric button. The Add Fabric screen comes up. The fields are:
Fabric Name - Enter the name of the fabric.
Fabric Template - This field has template options for creating specific types of fabric. Choose MSD_Fabric. The MSD screen comes up.
The fields in the screen are explained:
In the General tab, all fields are autopopulated with data. The fields consist of the Layer 2 and Layer 3 VXLAN segment identifier range, the default network and VRF templates, and the anycast gateway MAC address. Update the relevant fields as needed.
Layer 2 VXLAN VNI Range - Layer 2 VXLAN segment identifier range.
Layer 3 VXLAN VNI Range - Layer 3 VXLAN segment identifier range.
VRF Template - Default VRF template.
Network Template - Default network template.
VRF Extension Template - Default VRF extension template.
Network Extension Template - Default network extension template.
Anycast-Gateway-MAC - Anycast gateway MAC address.
Multisite Routing Loopback Id – The multicast routing loopback ID is populated in this field.
ToR Auto-deploy Flag - Select this check box to enable automatic deployment of the networks and VRFs in the Easy Fabric to the ToR switches in the External Fabric when you click Save & Deploy in the MSD fabric.
-
Click the DCI tab.
The fields are:
Multi-Site Overlay IFC Deploy Method – Choose how you will connect the data centers through the BGW, manually, in a back-to-back fashion or through a route server.
If you choose to connect them through a route server, you should enter the route server details.
Multi-Site Route Server List – Specify the IP addresses of the route server. If you specify more than one, separate the IP addresses by a comma.
Multi-Site Route Server BGP ASN List – Specify the BGP AS Number of the router server. If you specify more than one route server, separate the AS Numbers by a comma.
Multi-Site Underlay IFC Auto Deployment Flag - Check the check box to enable auto configuration. Uncheck the check box for manual configuration.
Delay Restore Time - Specifies the Multi-Site underlay and overlay control planes convergence time. The minimum value is 30 seconds and the maximum value is 1000 seconds.
Multi-Site CloudSec – Enables CloudSec configurations on border gateways. If you enable this field, the remaining three fields for CloudSec are editable. For more information, see Support for CloudSec in Multi-Site Deployment.
Enable Multi-Site eBGP Password - Enables eBGP password for Multi-Site underlay/overlay IFCs.
eBGP Password - Specifies the encrypted eBGP Password Hex String.
eBGP Authentication Key Encryption Type - Specifies the BGP key encryption type. It is 3 for 3DES and 7 for Cisco.
-
Click the Resources tab.
MultiSite Routing Loopback IP Range – Specify the Multi-Site loopback IP address range used for the EVPN Multi-Site function.
A unique loopback IP address is assigned from this range to each member fabric because each member site must have a Loopback 100 IP address assigned for overlay network reachability. The per-fabric loopback IP address is assigned on all the BGWs in a specific member fabric.
DCI Subnet IP Range and Subnet Target Mask – Specify the Data Center Interconnect (DCI) subnet IP address and mask.
-
Click the Configuration Backup tab.
Scheduled Fabric Backup: Check the check box to enable a daily backup. This backup tracks changes in running configurations on the fabric devices that are not tracked by configuration compliance.
Scheduled Time: Specify the scheduled backup time in a 24-hour format. This field is enabled if you check the Scheduled Fabric Backup check box.
Select both the check boxes to enable both back up processes.
The backup process is initiated after you click Save.
The backup configuration files are stored in the following path in DCNM: /usr/local/cisco/dcm/dcnm/data/archive
-
Click Save.
A message appears briefly at the bottom right part of the screen, indicating that you have created a new MSD fabric. After fabric creation, the fabric page comes up. The fabric name MSD-Parent-Fabric appears at the top left part of the screen.
Note
From Cisco DCNM Release 11.4(1), when you update the MSD fabric settings, only switches with roles relevant to MSD are updated.
Since the MSD fabric is a container, you cannot add a switch to it. The Add Switches button that is available in the Actions panel for member and standalone fabrics is not available for the MSD fabric.
When a new MSD is created, the newly created MSD fabric instance appears (as a rectangular box) on the Fabric Builder page. To go to the Fabric Builder page, click the ← button at the top left part of the MSD-Parent-Fabric page.
An MSD fabric is displayed as MSD in the Type field, and it contains the member fabric names in the Member Fabrics field. When no member fabric is created, None is displayed.
The steps for creation of an MSD fabric and moving member fabrics under it are:
-
Create an MSD fabric.
-
Create a new standalone fabric and move it under the MSD fabric as a member fabric.
Step 1 is completed. Step 2 is explained in the next section.
Creating and Moving a New Fabric Under the MSD Fabric as a Member
A new fabric is created as a standalone fabric. After you create a new fabric, you can move it under an MSD as a member. As a best practice, when you create a new fabric that is a potential member fabric (of an MSD), do not add networks and VRFs to the fabric. Move the fabric under the MSD and then add networks and VRFs for the MSD. That way, there will not be any need for validation (or conflict resolution) between the member and MSD fabric network and VRF parameters.
New fabric creation is explained in the Easy Fabric creation process. In the MSD document, fabric movement is covered. However, some pointers about a standalone (potential member) fabric:
The values that are displayed in the screen are automatically generated. The VXLAN VNI ID ranges (in the L2 Segment ID Range and L3 Partition ID Range fields) allocated for new network and VRF creation are values from the MSD fabric segment ID range. If you want to update the VXLAN VNI ranges or the VRF and Network VLAN ranges, ensure the following:
-
If you update a range of values, ensure that it does not overlap with other ranges.
-
You must update one range of values at a time. If you want to update more than one range of values, do it in separate instances. For example, if you want to update L2 and L3 ranges, you should do the following:
-
Update the L2 range and click Save.
-
Click the Edit Fabric option again, update the L3 range and click Save.
-
Ensure that the Anycast Gateway MAC, the Network Template and the VRF Template field values are the same as the MSD fabric. Else, member fabric movement to the MSD fail.
Other pointers:
-
Ensure that the Anycast Gateway MAC, the Network Template and the VRF Template field values are the same as the MSD fabric. Else, member fabric movement to the MSD fail.
-
The member fabric should have a Site ID configured and the Site ID must be unique among the members.
-
The BGP AS number should be unique for a member fabric.
-
The underlay subnet range for loopback0 should be unique.
-
The underlay subnet range for loopback1 should be unique.
After you click Save, a note appears at the bottom right part of the screen indicating that the fabric is created. When a fabric is created, the fabric page comes up. The fabric name appears at the top left part of the screen.
Simultaneously, the Fabric Builder page also displays the newly created fabric, Member1.
Simultaneously, the Fabric Builder page also displays the newly created fabric, Member1.
Moving the Member1 Fabric Under MSD-Parent-Fabric
You should go to the MSD fabric page to associate a member fabric under it.
If you are on the Fabric Builder page, click within the MSD-Parent-Fabric box to go to the MSD-Parent-Fabric page.
[If you are in the Member1 fabric page, you should go to the MSD-Parent-Fabrics-Docs fabric page. Click <- above the Actions panel. You will reach the Fabric Builder page. Click within the MSD-Parent-Fabric box].
-
In the MSD-Parent-Fabric page, go to the Actions panel and click Move Fabrics.
The Move Fabric screen comes up. It contains a list of fabrics.
Member fabrics of other MSD container fabrics are not displayed here.
The Member1 fabric is still a standalone fabric. A fabric is considered a member fabric of an MSD fabric only when you associate it with the MSD fabric. Also, each standalone fabric is a candidate for being an MSD fabric member, until you associate it to one of the MSD fabrics.
-
Since Member1 fabric is to be associated with the MSD fabric, select the Member1 radio button. The Add button is enabled.
-
Click Add.
Immediately, a message appears at the top of the screen indicating that the Member1 fabric is now associated with the MSD fabric MSD-Parent-Fabric. Now, the MSD-Parent-Fabric fabric page appears again.
-
Click the Move Fabrics option to check the fabric status. You can see that the fabric status has changed from standalone to member.
-
Close this screen.
-
Click ← above the Actions panel to go to the Fabric Builder page.
You can see that Member1 is now added to MSD fabric and is displayed in the Member Fabrics field.
MSD Fabric Topology View Pointers
-
MSD fabric topology view - Member fabrics and their switches are displayed. A boundary defines each member fabric. All fabric devices of the fabric are confined to the boundary.
All links are displayed, including intra-fabric links and Multi-Site (underlay and overlay), and VRF Lite links to remote fabrics.
-
Member fabric topology view - A member fabric and its switches are displayed. In addition, the connected external fabric is displayed.
-
A boundary defines a standalone VXLAN fabric, and each member fabric in an MSD fabric. A fabric’s devices are confined to the fabric boundary. You can move a switch icon by dragging it. For a better user experience, in addition to switches, DNCM 11.2(1) release allows you to move an entire fabric. To move a fabric, place the cursor within the fabric boundary (but not on a switch icon), and drag it in the desired direction.
Adding and Editing Links
To add a link, right-click anywhere in the topology and use the Add Link option. To edit a link, right-click on the link and use the Edit Link option.
Alternatively, you can use the Tabular view option in the Actions panel.
To know how to add links between border switches of different fabrics (inter-fabric links) or between switches in the same fabric (intra-fabric links), refer the Fabric Links topic.
Creating and Deploying Networks and VRFs in an MSD Fabric
In standalone fabrics, networks and VRFs are created for each fabric. In an MSD fabric, networks and VRFs should be created at the MSD fabric level. The networks and VRFs are inherited by all the member networks. You cannot create or delete networks and VRFs for member fabrics. However, you can edit them.
For example, consider an MSD fabric with two member fabrics. If you create three networks in the MSD fabric, then all three networks will automatically be available for deployment in both the member fabrics.
Though member fabrics inherit the MSD fabric's networks and VRFs, you have to deploy the networks and VRFs distinctly, for each fabric.
In DCNM 11.1(1) release, a deployment view is introduced for the MSD, in addition to the per-fabric deployment view. In this view, you can view and provision overlay networks for all member fabrics within the MSD, at once. However, you still have to apply and save network and VRF configurations distinctly, for each fabric.
Note |
Networks and VRFs are the common identifiers (represented across member fabrics) that servers (or end hosts) are grouped under so that traffic can be sent between the end hosts based on the network and VRF IDs, whether they reside in the same or different fabrics. Since they have common representation across member fabrics, networks and VRFs can be provisioned at one go. As the switches in different fabrics are physically and logically distinct, you have to deploy the same networks and VRFs separately for each fabric. |
For example, if you create networks 30000 and 30001 for an MSD that contains two member fabrics, the networks are automatically created for the member fabrics and are available for deployment.
In DCNM 11.1(1) release, you can deploy 30000 and 30001 on the border devices of all member fabrics through a single (MSD fabric) deployment screen. Prior to this, you had to access the first member fabric deployment screen, deploy 30000 and 300001 on the fabric's border devices, and then access the second member fabric deployment screen and deploy again.
Networks and VRFs are created in the MSD and deployed in the member fabrics. The steps are explained below:
-
Create networks and VRFs in the MSD fabric.
-
Deploy the networks and VRFs in the member fabric devices, one fabric at a time.
Creating Networks in the MSD Fabric
-
Click Control > Networks (under Fabrics submenu).
The Networks screen comes up.
-
Choose the correct fabric from SCOPE. When you select a fabric, the Networks screen refreshes and lists networks of the selected fabric.
-
Select MSD-Parent-Fabric from the list and click Continue at the top right part of the screen.
The Networks page comes up. This lists the list of networks created for the MSD fabric. Initially, this screen has no entries.
-
Click the + button at the top left part of the screen (under Networks) to add networks to the MSD fabric. The Create Network screen comes up. Most of the fields are autopopulated.
The fields in this screen are:
Network ID and Network Name - Specifies the Layer 2 VNI and name of the network. The network name should not contain any white spaces or special characters except underscore (_) and hyphen (-).
VRF Name - Allows you to select the Virtual Routing and Forwarding (VRF).
When no VRF is created, this field is blank. If you want to create a new VRF, click the + button. The VRF name should not contain any white spaces or special characters except underscore (_), hyphen (-), and colon (:).
Note
You can also create a VRF by clicking the VRF View button on the Networks page.
Layer 2 Only - Specifies whether the network is Layer 2 only.
Network Template - Allows you to select a network template.
Network Extension Template - This template allows you to extend the network between member fabrics.
VLAN ID - Specifies the corresponding tenant VLAN ID for the network.
Network Profile section contains the General and Advanced tabs, explained below.
General tab
IPv4 Gateway/NetMask - Specifies the IPv4 address with subnet.
IPv6 Gateway/Prefix - Specifies the IPv6 address with subnet.
VLAN Name - Enter the VLAN name.
If the VLAN is mapped to more than one subnet, enter the anycast gateway IP addresses for those subnets.
Interface Description - Specifies the description for the interface.
MTU for the L3 interface - Enter the MTU for Layer 3 interfaces.
IPv4 Secondary GW1 - Enter the gateway IP address for the additional subnet.
IPv4 Secondary GW2 - Enter the gateway IP address for the additional subnet.
Advanced tab - Optionally, specify the advanced profile settings by clicking the Advanced tab. The options are:
-
ARP Suppression
-
DHCPv4 Server 1 and DHCPv4 Server 2 - Enter the DHCP relay IP address of the first and second DHCP servers.
-
DHCPv4 Server VRF - Enter the DHCP server VRF ID.
-
Loopback ID for DHCP Relay interface - Enter the loopback ID of the DHCP relay interface.
-
Routing Tag – The routing tag is autopopulated. This tag is associated with each gateway IP address prefix.
-
TRM enable – Select the check box to enable TRM.
For more information, see Overview of Tenant Routed Multicast.
-
L2 VNI Route-Target Both Enable - Select the check box to enable automatic importing and exporting of route targets for all L2 virtual networks.
Note
From Cisco DCNM Release 11.5(1), the Enable L3 Gateway on Border field is not available as part of the MSD network settings. You can enable a Layer 3 gateway on the border switches at a fabric level. For more information, see Creating Networks for the Standalone Fabric.
In the MSD fabric level, if the Enable L3 Gateway on Border check box is selected and you are upgrading to Cisco DCNM Release 11.5(1), then it is automatically removed from the MSD fabric level during upgrade.
-
A sample of the Create Network screen:
-
-
Click Create Network. A message appears at the bottom right part of the screen indicating that the network is created. The new network (MyNetwork_30000) appears on the Networks page that comes up.
Editing Networks in the MSD Fabric
-
In the Networks screen of the MSD fabric, select the network you want to edit and click the Edit icon at the top left part of the screen.
The Edit Network screen comes up.
You can edit the Network Profile part (General and Advanced tabs) of the MSD fabric network.
-
Click Save at the bottom right part of the screen to save the updates.
Network Inheritance from MSD-Parent-Fabric to Member1
MSD-Parent-Fabric fabric contains one member fabric, Member1. Go to the Select a Fabric page to access the Member1 fabric.
-
Click Control > Networks (under Fabrics submenu).
The Networks screen comes up.
-
Choose the correct fabric from SCOPE. When you select a fabric, the Networks screen refreshes and lists networks of the selected fabric.
Editing Networks in the Member Fabric
An MSD can contain multiple fabrics. These fabrics forward BUM traffic via Multicast or Ingress replication. Even if all the fabrics use multicast for BUM traffic, the multicast groups within these fabrics need not be the same.
When you create a network in MSD, it is inherited by all the member fabrics. However, the multicast group address is a fabric instance variable. To edit the multicast group address, you need to navigate to the member fabric and edit the network. For more information about the Multicast Group Address field, see Creating Networks for the Standalone Fabric.
-
Select the network and click the Edit option at the top left part of the window. The Edit Network window comes up.
-
Update the multicast group address in one of the following ways:
-
Under Network Profile, click the Generate Multicast IP button to generate a new multicast group address for the selected network, and click Save.
-
Click the Advanced tab in the Network Profile section, update the multicast group address, and click Save.
-
Note |
The Generate Multicast IP option is only available for member fabric networks and not MSD networks. |
Deleting Networks in the MSD and Member Fabrics
You can only delete networks from the MSD fabric, and not member fabrics. To delete networks and corresponding VRFs in the MSD fabric, follow this order:
-
Undeploy the networks on the respective fabric devices before deletion.
-
Delete the networks from the MSD fabric. To delete networks, use the delete (X) option at the top left part of the Networks screen. You can delete multiple networks at once.
Note
When you delete networks from the MSD fabric, the networks are automatically removed from the member fabrics too.
-
Undeploy the VRFs on the respective fabric devices before deletion.
-
Delete the VRFs from the MSD fabric by using the delete (X) option at the top left part of the screen. You can delete multiple VRF instances at once.
Creating VRFs in the MSD Fabric
-
From the MSD fabric's Networks page, click the VRF View button at the top right part of the screen to create VRFs.
-
Choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists VRFs of the selected fabric.
-
Choose the MSD fabric (MSD-Parent-Fabric) from the drop-down box and click Continue. The Networks page comes up.
-
Click VRF View at the top right part of the Networks page].
The VRFs page comes up. This lists the list of VRFs created for the MSD fabric. Initially, this screen has no entries.
-
-
Click the + button at the top left part of the screen to add VRFs to the MSD fabric. The Create VRF screen comes up. Most of the fields are autopopulated.
The fields in this screen are:
VRF ID and VRF Name - The ID and name of the VRF.
The VRF ID is the VRF VNI or the L3 VNI of the tenant.
Note
For ease of use, the VRF creation option is also available while you create a network.
VRF Template - This is populated with the Default_VRF template.
VRF Extension Template - This template allows you to extend the VRF between member fabrics.
-
General tab – Enter the VLAN ID of the VLAN associated with the VRF, the corresponding Layer 3 virtual interface, and the VRF ID.
-
Advanced tab
Routing Tag – If a VLAN is associated with multiple subnets, then this tag is associated with the IP prefix of each subnet. Note that this routing tag is associated with overlay network creation too.
Redistribute Direct Route Map – Specifies the route map name for redistribution of routes in the VRF.
Max BGP Paths and Max iBGP Paths – Specifies the maximum BGP and iBGP paths.
TRM Enable – Select the check box to enable TRM.
If you enable TRM, then the RP address, and the underlay multicast address must be entered.
For more information, see Overview of Tenant Routed Multicast.
Is RP External – Enable this checkbox if the RP is external to the fabric. If this field is unchecked, RP is distributed in every VTEP.
RP Address – Specifies the IP address of the RP.
RP Loopback ID – Specifies the loopback ID of the RP, if Is RP External is not enabled.
Underlay Multicast Address – Specifies the multicast address associated with the VRF. The multicast address is used for transporting multicast traffic in the fabric underlay.Note
The multicast address in the Default MDT Address for TRM VRFs field in the fabric settings screen is auto-populated in this field. You can override this field if a different multicast group address should be used for this VRF.
Overlay Multicast Groups – Specifies the multicast group subnet for the specified RP. The value is the group range in “ip pim rp-address” command. If the field is empty, 224.0.0.0/24 is used as default.
Enable IPv6 link-local Option - Select the check box to enable the IPv6 link-local option under the VRF SVI. If this check box is unchecked, IPv6 forward is enabled.
Advertise Host Routes - Select the checkbox to control advertisement of /32 and /128 routes to Edge Routers.
Advertise Default Route - Select the checkbox to control advertisement of default routes within the fabric.
A sample screenshot:
Advanced tab:
-
Click Create VRF.
The MyVRF_50000 VRF is created and appears on the VRFs page.
Editing VRFs in the MSD Fabric
-
In the VRFs screen of the MSD fabric, select the VRF you want to edit and click the Edit icon at the top left part of the screen.
The Edit VRF screen comes up.
You can edit the VRF Profile part (General and Advanced tabs).
-
Click Save at the bottom right part of the screen to save the updates.
VRF Inheritance from MSD-Parent-Fabric to Member1
MSD-Parent-Fabric contains one member fabric, Member1. Do the following to access the member fabric page.
-
Choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists VRFs of the selected fabric.
-
Click the VRF View button. On the VRFs page, you can see that the VRF created for the MSD is inherited to its member.
Deleting VRFs in the MSD and Member Fabrics
You can only delete networks from the MSD fabric, and not member fabrics. To delete networks and corresponding VRFs in the MSD fabric, follow this order:
-
Undeploy the networks on the respective fabric devices before deletion.
-
Delete the networks from the MSD fabric.
-
Undeploy the VRFs on the respective fabric devices before deletion.
-
Delete the VRFs from the MSD fabric by using the delete (X) option at the top left part of the screen. You can delete multiple VRF instances at once.
Note
When you delete VRFs from the MSD fabric, they are automatically removed from the member fabrics too.
Editing VRFs in the Member Fabric
You cannot edit VRF parameters at the member fabric level. Update VRF settings in the MSD fabric. All member fabrics are automatically updated.
Deleting VRFs in the Member Fabric
You cannot delete VRFs at the member fabric level. Delete VRFs in the MSD fabric. The deleted VRFs are automatically removed from all member fabrics.
Step 1 of the following is explained. Step 2 information is mentioned in the next subsection.
-
Create networks and VRFs in the MSD fabric.
-
Deploy the networks and VRFs in the member fabric devices, one fabric at a time.
Deployment and Undeployment of Networks and VRFs in Member Fabrics
Before you begin, ensure that you have created networks at the MSD fabric level since the member fabric inherits networks and VRFs created for the MSD fabric.
Note |
The deployment (and undeployment) of networks and VRFs in member fabrics are the same as explained for standalone fabrics. Refer Creating and Deploying Networks and VRFs. |
Support for CloudSec in Multi-Site Deployment
CloudSec feature allows secured data center interconnect in a multi-site deployment by supporting source-to-destination packet encryption between border gateway devices in different fabrics.
CloudSec feature is supported on Cisco Nexus 9000 Series FX2 platform with Cisco NX-OS Release 9.3(5) or later. The border gateways, border gateway spines, and border gateway superspines that are FX2 platforms, and run Cisco NX-OS Release 9.3(5) or later are referred as CloudSec capable switches.
Cisco DCNM Release 11.4(1) provides an option to enable CloudSec in an MSD fabric.
Note |
The CloudSec session is point to point over DCI between border gateways (BGWs) on two different sites. All communication between sites uses Multi-Site PIP instead of VIP. Enabling CloudSec requires a switch from VIP to PIP, which could cause traffic disruption for data flowing between sites. Therefore, it is recommended to enable or disable CloudSec during a maintenance window. |
You can also watch the video that demonstrates how to configure the CloudSec feature. See Video: Configuring CloudSec in Cisco DCNM.
Enabling CloudSec in MSD
Navigate to Control > Fabrics > Fabric Builder. You can either create a new MSD fabric by clicking Create Fabric or edit the existing MSD fabric by clicking Edit Fabric.
Under the DCI tab, you can specify the CloudSec configuration details.
Multi-Site CloudSec – Enables CloudSec configurations on border gateways. If you enable this field, the remaining CloudSec fields are editable.
Multi-Site CloudSec – Enables CloudSec configurations on border gateways. If you enable this field, the remaining three fields for CloudSec are editable.
When Cloudsec is enabled at MSD level, DCNM also enables dci-advertise-pip under evpn multisite border-gateway and tunnel-encryption on the uplinks for all Cloudsec capable gateways.
When you click Save & Deploy, you can verify theses configs in the Preview Config window for the border gateway switches.
Note – CloudSec isn’t supported if the border gateway has vPC or TRM is enabled on it, that is, TRM enabled on multisite overlay IFC. If CloudSec is enabled in this scenario, appropriate warning or error messages are generated.
CloudSec Key String – Specifies the hex key string. Enter a 66 hexadecimal string if you choose AES_128_CMAC or enter a 130 hexadecimal string if you choose AES_256_CMAC.
CloudSec Cryptographic Algorithm – Choose AES_128_CMAC or AES_256_CMAC.
CloudSec Enforcement – Specifies whether the CloudSec enforcement should be strict or loose.
strict – Deploys the CloudSec configuration to all the border gateways in fabrics in MSD. If there are any border gateways that don’t support CloudSec, then an error message is generated, and the configuration isn’t pushed to any switch.
If strict is chosen, the tunnel-encryption must-secure CLI is pushed to the CloudSec enabled gateways within MSD.
loose – Deploys the CloudSec configuration to all the border gateways in fabrics in MSD. If there are any border gateways that don’t support CloudSec, then a warning message is generated. In this case, the CloudSec config is only deployed to the switches that support CloudSec. If loose is chosen, the tunnel-encryption must-secure CLI is removed if it exists.
Note |
There should be at least two fabrics in MSD with border gateways that support CloudSec. If there’s only one fabric with a CloudSec capable device, then the following error message is generated: CloudSec needs to have at least 2 sites that can support CloudSec. |
To remove this error, meet the criteria of having at least two sites that can support CloudSec or disable CloudSec.
CloudSec Status Report Timer – Specifies the CloudSec Operational Status periodic report timer in minutes. This value specifies how often the DCNM polls the CloudSec status data from the switch. The default value is 5 minutes and the range is from 5 to 60 minutes.
Using the CloudSec feature in DCNM, you can have all the gateways within the MSD to use the same keychain (and have only one key string) and policy. You can provide one key chain string for DCNM to form the key chain policy. DCNM forms the encryption-policy by taking all default values. DCNM pushes the same key chain policy, the same encryption-policy, and encryption-peer policies to each CloudSec capable gateways. On each gateway, there’s one encryption-peer policy for each remote gateway that is CloudSec capable, using the same keychain and same key policy.
If you don’t want to use the same key for the whole MSD fabric or want to enable CloudSec on a subset of all sites, you can use switch_freeform to manually push the CloudSec config to the switches.
Capture all the CloudSec config in switch_freeform.
For example, the below config is included in the switch_freeform policy:
feature tunnel-encryption
evpn multisite border-gateway 600
dci-advertise-pip
tunnel-encryption must-secure-policy
tunnel-encryption policy CloudSec_Policy1
tunnel-encryption source-interface loopback20
key chain CloudSec_Key_Chain1 tunnel-encryption
key 1000
key-octet-string 7 075e731f1a5c4f524f43595f507f7d73706267714752405459070b0b0701585440 cryptographic-algorithm AES_128_CMA
tunnel-encryption peer-ip 192.168.0.6
keychain CloudSec_Key_Chain1 policy CloudSec_Policy1
Add tunnel-encryption in the Freeform Config of the uplink interface policy which will generate config like the following:
interface ethernet1/13
no switchport
ip address 192.168.1.14/24 tag 54321
evpn multisite dci-tracking
tunnel-encryption
mtu 9216
no shutdown
For more information, see Enabling Freeform Configurations on Fabric Switches.
When CloudSec configuration is added to or removed from the switch, the DCI uplinks will flap, which will trigger multisite BGP session flapping. For multisite with existing cross site traffic, there will be traffic disruption during this transition. Therefore, it is recommended to make the transition during a maintenance window.
If you’re migrating an MSD fabric with the CloudSec configuration into DCNM, the Cloudsec related configuration is captured in switch_freeform and interface freeform config. You do not need to turn on Multi-Site Cloudsec in the MSD fabric setting. If you want to add more fabrics and establish CloudSec tunnels which share the same CloudSec policy including key as the existing one, then you can enable the CloudSec config in the MSD fabric settings. The CloudSec parameters in the MSD fabric setting need to match the existing CloudSec configuration on the switch. The CloudSec configuration is already captured in the freeform config, and enabling CloudSec in MSD will also generate config intents. Therefore, there’s a double intent. For example, if you want to change the CloudSec key in the MSD settings, you need to remove the CloudSec freeform config because DCNM won’t modify config in switch_freeform. Otherwise, the key in the MSD fabric settings is a conflict with the key in the freeform config.
Viewing CloudSec Operational State
From Cisco DCNM 11.5(1), you can use CloudSec Operational View to check the operational status of the CloudSec sessions if CloudSec is enabled on the MSD fabric.
Procedure
Step 1 |
Choose an MSD fabric. The fabric topology window appears. |
||||||||||||||
Step 2 |
Click Tabular view in the Actions pane. |
||||||||||||||
Step 3 |
Choose the CloudSec Operational View tab. |
||||||||||||||
Step 4 |
If CloudSec is disabled, the CloudSec Operational View tab isn’t displayed. The Operational View tab has the following fields and descriptions.
All these columns are sortable.
|
Troubleshooting a CloudSec Session
If a CloudSec session is down, you can find more information about it using Programmable Report.
Procedure
Step 1 |
Navigate to Applications > Programmable report. |
Step 2 |
Click the Create Report icon. |
Step 3 |
Specify a report name, select the MSD fabric on which the report job should be run, and click Next. |
Step 4 |
From the Template drop-down list, select fabric_cloudsec_oper_status and click Create Job. The status will change to a green tick indicating Success after the report has been successfully generated. |
Step 5 |
Click the report to view it. This report is similar to the CloudSec Operation View tab. |
Step 6 |
Click View Details to view more information about the CloudSec session status. |
Step 7 |
Click the operational status for a session to view the detailed info about the CloudSec session for each peer fabric and device. |
Removing a Fabric From an MSD
To remove a fabric from an MSD fabric, perform the following steps:
Before you begin
Note |
From Cisco DCNM Release 11.4(1), after removing an individual fabric from MSD, underlay and overlay IFCs are deleted. If IFCs are extended, an error is reported to disallow the fabric remove. |
Procedure
Step 1 |
From the Fabric Builder window, click an MSD fabric. |
Step 2 |
Click Move Fabric in the Actions menu. |
Step 3 |
In the Move Fabric window, select the respective radio button of the fabric that you want to remove and click Remove. In the fabric removal notification window, click Close. |
Step 4 |
Click Save & Deploy for the MSD in the Fabric Builder window. |
Step 5 |
Click Deploy Config in the Config Deployment window. Click Close. |
Step 6 |
Navigate to the fabric that you removed from MSD and click Save & Deploy. |
Step 7 |
Click Deploy Config in the Config Deployment window. Click Close. |
Moving a Standalone Fabric (With Existing Networks and VRFs) to an MSD Fabric
If you move a standalone fabric with existing networks and VRFs to an MSD fabric as a member, ensure that common networks (that is, L2 VNI and L3 VNI information), anycast gateway MAC, and VRF and network templates are the same across the fabric and the MSD. DCNM validates the standalone fabric (network and VRF information) against the (network and VRF information) of the MSD fabric to avoid duplicate entries. An example of duplicate entries is two common network names with a different network ID. After validation for any conflicts, the standalone fabric is moved to the MSD fabric as a member fabric. Details:
-
The MSD fabric inherits the networks and VRFs of the standalone fabric that do not exist in the MSD fabric. These networks and VRFs are in turn inherited by the member fabrics.
-
The newly created member fabric inherits the networks and VRFs of the MSD fabric (that do not exist in the newly created member fabric).
-
If there are conflicts between the standalone and MSD fabrics, validation ensures that an error message is displayed. After the updation, when you move the member fabric to the MSD fabric, the move will be successful. A message comes up at the top of the page indicating that the move is successful.
If you move back a member fabric to standalone status, then the networks and VRFs remain as they are, but they remain relevant as in an independent fabric, outside the purview of an MSD fabric.
Managing Switches Using LAN Classic Templates
From Cisco DCNM Release 11.4(1), you can use the LAN_Classic and Fabric_Group templates to manage the switches that you used to previously manage in the DCNM Classic LAN deployment.
The LAN_Classic fabric template is a generic fabric template to manage Cisco Nexus switches.
Guidelines and Limitations
-
Fabrics using the LAN_Classic fabric template can be changed to use the External_Fabric_11_1 fabric template and then use all its associated functionalities. Note that this is the only supported fabric template conversion and it’s nonreversible.
-
The LAN_Classic fabric can be added as a member of an MSD fabric.
-
Only Cisco Nexus switches are supported in the LAN_Classic fabric.
-
The TOR Auto-Deploy functionality is supported in the LAN_Classic member fabric when a switch with the ToR role is in the fabric. For more information, see Configuring ToR Switches and Deploying Networks.
-
If you are using the Cisco Nexus 7000 Series Switch with Cisco NX-OS Release 6.2(24a) on the LAN Classic or External fabrics, make sure to enable AAA IP Authorization in the fabric settings.
-
The following features in the LAN_Classic template provide the same support as they do for the External_Fabric_11_1 template:
The following features are supported:
-
Configuration compliance
-
Backup or restore of fabric
-
Network Insights
-
Performance monitoring
-
VMM
-
Topology view
-
Kubernetes visualization
-
RBAC
For more information, refer to the feature specific sections.
-
Creating a LAN Classic Fabric
Procedure
Step 1 |
Navigate to Control > Fabrics > Fabric Builder. |
||||
Step 2 |
Click Create Fabric. |
||||
Step 3 |
Enter the fabric name and choose LAN_Classic from Fabric Template drop - down list. |
||||
Step 4 |
The General tab is displayed by default. The field in this tab is: Fabric Monitor Mode – Uncheck the check box if you want DCNM to manage the fabric. Keep the check box selected to enable only monitoring of the fabric. In this state, you can’t deploy configurations on its switches. The configurations must be pushed for devices before you discover them in the fabric. You can’t push configurations in the monitor mode. |
||||
Step 5 |
Click Advanced tab. The fields in this tab are: vPC Peer Link VLAN - The vPC peer link VLAN ID is autopopulated. Update the field to reflect the correct value. Power Supply Mode - Choose the appropriate power supply mode. Enable MPLS Handoff - Select the check box to enable the MPLS Handoff feature. For more information, see the Border Provisioning Use Case in VXLAN BGP EVPN Fabrics - MPLS SR and LDP Handoff chapter. Underlay MPLS Loopback Id: Specifies the underlay MPLS loopback ID. The default value is 101. Enable AAA IP Authorization - Enables AAA IP authorization, when IP Authorization is enabled in the AAA Server. Enable DCNM as Trap Host - Select this check box to enable DCNM as a trap host. Enable CDP for Bootstrapped Switch - Enables CDP on management interface. Enable NX-API - Specifies enabling of NX-API. This check box is unchecked by default. Enable NX-API on HTTP port - Specifies enabling of NX-API on HTTP. This check box is unchecked by default. Enable this check box and the Enable NX-API check box to use HTTP. If you uncheck this check box, the applications that use NX-API and supported by Cisco DCNM, such as Layer 4-Layer 7 services (L4-L7 services), VXLAN OAM, and so on, start using the HTTPS instead of HTTP.
Inband Mgmt: For External and Classic LAN Fabrics, this knob enables DCNM to import and manage of switches with inband connectivity (reachable over switch loopback, routed, or SVI interfaces) , in addition to management of switches with out-of-band connectivity (aka reachable over switch mgmt0 interface). The only requirement is that for Inband managed switches, there should be IP reachability from DCNM to the switches via the eth2 aka inband interface. For this purpose, static routes may be needed on the DCNM, that in turn can be configured via the Administration->Customization->Network Preferences option. After enabling Inband management, during discovery, provide the IPs of all the switches to be imported using Inband Management and set maximum hops to 0. DCNM has a pre-check that validates that the Inband managed switch IPs are reachable over the eth2 interface. Once the pre-check has passed, DCNM then discovers and learns about the interface on that switch that has the specified discovery IP in addition to the VRF that the interface belongs to. As part of the process of switch import/discovery, this information is captured in the baseline intent that is populated on the DCNM. For more information, see Inband Management in External Fabrics and LAN Classic Fabrics.
Enable Precision Time Protocol (PTP): Enables PTP across a fabric. When you select this check box, PTP is enabled globally and on core-facing interfaces. Additionally, the PTP Source Loopback Id and PTP Domain Id fields are editable. For more information, see Precision Time Protocol for External Fabrics and LAN Classic Fabrics. Fabric Freeform - You can apply configurations globally across all the devices discovered in the external fabric using this freeform field. AAA Freeform Config – Specifies the AAA freeform configs. |
||||
Step 6 |
Click the Resources tab. The fields in this tab are: Subinterface Dot1q Range - The subinterface 802.1Q range and the underlay routing loopback IP address range are autopopulated. Underlay Routing Loopback IP Range - Specifies loopback IP addresses for the protocol peering. Underlay MPLS Loopback IP Range - Specifies the underlay MPLS SR or LDP loopback IP address range. The IP range should be unique, that is, it shouldn’t overlap with IP ranges of the other fabrics. |
||||
Step 7 |
Click Configuration tab. The fields in this tab are: Hourly Fabric Backup: Select the check box to enable an hourly backup of fabric configurations and the intent. Scheduled Fabric Backup: Check the check box to enable a daily backup. Scheduled Time: Specify the scheduled backup time in a 24-hour format. This field is enabled if you check the Scheduled Fabric Backup check box.
The backup and restore process is similar to that of an external fabric. For more information about backing up and restoring external fabrics, see Fabric Backup and Restore. |
||||
Step 8 |
Click Bootstrap tab. The fields in this tab are: Enable Bootstrap (For NX-OS Switches Only) - Select this check box to enable the bootstrap feature for only Cisco Nexus switches. After you enable bootstrap, you can enable the DHCP server for automatic IP address assignment using one of the following methods:
Enable Local DHCP Server - Select this check box to initiate enabling of automatic IP address assignment through the local DHCP server. When you select this check box, all the remaining fields become editable. DHCP Version – Select DHCPv4 or DHCPv6 from this drop-down list. When you select DHCPv4, the Switch Mgmt IPv6 Subnet Prefix field is disabled. If you select DHCPv6, the Switch Mgmt IP Subnet Prefix is disabled.
If you don’t select this check box, DCNM uses the remote or external DHCP server for automatic IP address assignment. DHCP Scope Start Address and DHCP Scope End Address - Specifies the first and last IP addresses of the IP address range to be used for the switch out of band POAP. Switch Mgmt Default Gateway - Specifies the default gateway for the management VRF on the switch. Switch Mgmt IP Subnet Prefix - Specifies the prefix for the Mgmt0 interface on the switch. The prefix should be between 8 and 30. DHCP scope and management default gateway IP address specification - If you specify the management default gateway IP address 10.0.1.1 and subnet mask 24, ensure that the DHCP scope is within the specified subnet, between 10.0.1.2 and 10.0.1.254. Switch Mgmt IPv6 Subnet Prefix - Specifies the IPv6 prefix for the Mgmt0 interface on the switch. The prefix should be between 112 and 126. This field is editable if you enable IPv6 for DHCP. Enable AAA Config - Enables AAA configure. It includes AAA configs from the Advanced tab during device bootup. Bootstrap Freeform Config - (Optional) Enter extra commands as needed. For example, if you’re using AAA or remote authentication-related configurations, add these configurations in this field to save the intent. After the devices boot up, they contain the intent defined in the Bootstrap Freeform Config field. Copy-paste the running-config to a freeform config field with correct indentation, as seen in the running configuration on the NX-OS switches. The freeform config must match the running config. For more information, see Resolving Freeform ConfigErrors in Switches. DHCPv4/DHCPv6 Multi Subnet Scope - Specifies the field to enter one subnet scope per line. This field is editable after you check the Enable Local DHCP Server check box. The format of the scope should be defined as: DHCP Scope Start Address, DHCP Scope End Address, Switch Management Default Gateway, Switch Management Subnet Prefix For example: 10.6.0.2, 10.6.0.9, 10.6.0.1, 24 After the fabric is created, the fabric topology page comes up. |
||||
Step 9 |
Click ThousandEyes Agent tab. This feature is supported on Cisco DCNM Release 11.5(3) only. For more information, refer to Configuring Global Settings for ThousandEyes Enterprise Agent. The fields on this tab are:
|
Adding Switches to LAN Classic Fabric
Procedure
Step 1 |
Click Add switches. The Inventory Management window comes up. You can also add switches by clicking Tabular View > Switches > + . |
Step 2 |
Enter IP address (Seed IP) of the switch. |
Step 3 |
Enter the administrator username and password of the switch. |
Step 4 |
Click Start discovery at the bottom part of the screen. The Scan Details section comes up shortly. Since the Max Hops field was populated with 2, the switch with the specified IP address and switches two hops from it are populated. |
Step 5 |
Select the check boxes next to the concerned switches and click Import into fabric. You can discover multiple switches at the same time. The switches must be properly cabled and connected to the DCNM server and the switch status must be manageable. The switch discovery process is initiated. The Progress column displays the progress. After DCNM discovers the switch, the screen closes and the fabric screen comes up again. The switch icons are seen at the centre of the fabric screen. |
Step 6 |
Click Refresh topology to view the latest topology view. For more information, see: |
Creating a Fabric Group and Associating Member Fabrics
This procedure shows how to a create a Fabric_Group and add LAN_Classic fabrics. The Fabric_Group template is used for grouping LAN_Classic fabrics for visualization.
The following functionalities aren't supported in a Fabric_Group:
-
Fabric backup or restore
-
VXLAN overlay or IFC deployment
-
Changing fabric template to and from any other fabric template
-
Since Fabric_Group doesn't manage any configurations, clicking Save & Deploy reports an error.
Procedure
Step 1 |
Navigate to Control > Fabrics > Fabric Builder. |
||
Step 2 |
Click Create Fabric. |
||
Step 3 |
Enter the fabric name and choose Fabric_Group from the Fabric Template drop - down list. |
||
Step 4 |
Click Save. |
||
Step 5 |
In the Actions panel, click Move Fabrics. |
||
Step 6 |
Select a LAN_Classic fabric in the Move Fabric window.
|
||
Step 7 |
Click Add. Similarly, you can remove a member fabric by selecting it and clicking Remove. |
Support for Inter-Fabric Connection in LAN Classic Fabric Template
The LAN_Classic fabric supports VRF-Lite, Multi-Site, and MPLS IFCs with these conditions:
-
The LAN_Classic fabric as a destination for DCI/VRF-Lite and Multi-Site IFCs is supported, but you can only manually create them by providing the required information. They won’t be automatically created even when the auto deployment options are enabled in the Easy_Fabric_11_1 and MSD_Fabric_11_1 fabrics.
-
You can’t add nonexistent (meta) switches to a LAN_Classic fabric. A meta switch is a placeholder for a switch or device that DCNM can’t discover.
-
The base BGP configurations for the 'Edge Router' and 'Core Router' switch roles aren’t auto generated. Configure them using the switch_freeform policies or other suitable means.
-
If MPLS Handoff is enabled in the fabric settings, MPLS base configurations are auto generated for the 'Edge Router' and 'Core Router' switch roles.
Inband Management in External Fabrics and LAN Classic Fabrics
From Release 11.5(1), Cisco DCNM allows you to import or discover switches with inband connectivity for External and LAN Classic fabrics in Brownfield deployments only. Enable inband management, per fabric, while configuring or editing the Fabric settings. You cannot import or discover switches with inband connectivity using POAP.
After configuration, the Fabric tries to discover switches based on the VRF of the inband management. The fabric template determines the VRF of inband switch using seed IP. If there are multiple VRFs for same seed IP, then no intent will be learnt for seed interfaces. You must create intent/configuration manually.
After configuring\editing the Fabric settings, you must Save and Deploy. You cannot change the Inband Mgmt settings after you import inband managed switches to the Fabric. If you uncheck the checkbox, the following error message is generated.
Inband IP <<IP Address>> cannot be used to import the switch,
please enable Inband Mgmt in fabric settings and retry.
After the switches are imported to the Fabric, you must manage the interfaces to create intent. Create the intent for the interfaces that you’re importing the switch. Edit\update the Interface configuration. When you try to change the Interface IP, for this inband managed switch, an error message is generated:
Interface <<interface_name>> is used as seed or next-hop egress interface
for switch import in inband mode.
IP/Netmask Length/VRF changes are not allowed for this interface.
<<switch-name>>: Mgmt0 IP Address (<ip-address>) cannot be changed,
when is it used as seed IP to discover the switch.
Create a policy for next-hop interfaces. Routes to DCNM from 3rd party devices may contain multiple interfaces, known as ECMP routes. Find the next-hop interface and create an intent for the switch. Interface IP and VRF changes are not allowed.
If inband management is enabled, during Image management, eth2 IP address is used to copy images on the switch, in ISSU, EPLD, RPM & SMU installations flows.
The fabric <<fabric name>> was updated with below message:
Fabric Settings cannot be changed for Inband Mgmt, when switches are already imported
using inband Ip. Please remove the existing switches imported using Inband Ip from the fabric,
then change the Fabric Setttings.
However, the same fabric can contain switches imported using both inband and out-of-band connectivity.
Precision Time Protocol for External Fabrics and LAN Classic Fabrics
From Release 11.5(1), in the fabric settings for the External_Fabric_11_1 or LAN_Classic template, select the Enable Precision Time Protocol (PTP) check box to enable PTP across a fabric. When you select this check box, PTP is enabled globally and on core-facing interfaces. Additionally, the PTP Loopback Id and PTP Domain Id fields are editable.
The PTP feature is supported with Cisco Nexus 9000 Series cloud-scale switches, with NX-OS version 7.0(3)I7(1) or later. Warnings are displayed if there are non-cloud scale devices in the fabric, and PTP is not enabled. Examples of the cloud-scale devices are Cisco Nexus 93180YC-EX, Cisco Nexus 93180YC-FX, Cisco Nexus 93240YC-FX2, and Cisco Nexus 93360YC-FX2 switches. For more information, refer to https://www.cisco.com/c/en/us/products/switches/nexus-9000-series-switches/index.html#~products.
Note |
PTP global configuration is supported with Cisco Nexus 3000 Series switches; however, PTP and ttag configurations are not supported. |
For more information, see the Configuring PTP chapter in Cisco Nexus 9000 Series NX-OS System Management Configuration Guide and Cisco Network Insights for Resources Application for Cisco DCNM User Guide.
For External and LAN Classic fabric deployments, you have to enable PTP globally, and also enable PTP on core-facing interfaces. The interfaces could be configured to the external PTP server like a VM or Linux-based machine. Therefore, the interface should be edited to have a connection with the grandmaster clock. For PTP and TTAG configurations to be operational on External and LAN Classic Fabrics, you must sync up of Switch Configs to DCNM using the host_port_resync policy. For more information, see Sync up Out-of-Band Switch Interface Configurations with DCNM.
It is recommended that the grandmaster clock should be configured outside of Easy Fabric and it is IP reachable. The interfaces toward the grandmaster clock need to be enabled with PTP via the interface freeform config.
All core-facing interfaces are auto-enabled with the PTP configuration after you click Save & Deploy. This action ensures that all devices are PTP synced to the grandmaster clock. Additionally, for any interfaces that are not core-facing, such as interfaces on the border devices and leafs that are connected to hosts, firewalls, service-nodes, or other routers, the ttag related CLI must be added. The ttag is added for all traffic entering the VXLAN EVPN fabric and the ttag must be stripped when traffic is exiting this fabric.
Here is the sample PTP configuration:feature ptp
feature ptp
ptp source 100.100.100.10 -> IP address of the loopback interface (loopback0)
that is already created, or user-created loopback interface in the fabric settings
ptp domain 1 -> PTP domain ID specified in fabric settings
interface Ethernet1/59 -> Core facing interface
ptp
interface Ethernet1/50 -> Host facing interface
ttag
ttag-strip
The following guidelines are applicable for PTP:
-
The PTP feature can be enabled in a fabric when all the switches in the fabric have Cisco NX-OS Release 7.0(3)I7(1) or a higher version. Otherwise, the following error message is displayed:
PTP feature can be enabled in the fabric, when all the switches have NX-OS Release 7.0(3)I7(1) or higher version. Please upgrade switches to NX-OS Release 7.0(3)I7(1) or higher version to enable PTP in this fabric.
-
For hardware telemetry support in NIR, the PTP configuration is a prerequisite.
-
If you are adding a non-cloud scale device to an existing fabric which contains PTP configuration, the following warning is displayed:
TTAG is enabled fabric wide, when all devices are cloud-scale switches so it cannot be enabled for newly added non cloud-scale device(s).
-
If a fabric contains both cloud-scale and non-cloud scale devices, the following warning is displayed when you try to enable PTP:
TTAG is enabled fabric wide when all devices are cloud-scale switches and is not enabled due to non cloud-scale device(s).
-
TTAG configuration is generated for all the devices if host configuration sync up is performed on all the devices. Ttag configuration will not be generated for any newly added devices if host configuration sync up is not performed on all newly added devices.
If the configuration is not synced, the following warning is displayed:
TTAG on interfaces with PTP feature can only be configured for cloud-scale devices. It will not be enabled on any newly added switches due to the presence of non cloud-scale devices.
-
PTP and TTAG configurations are deployed on host interfaces.
-
PTP and TTAG Configurations are supported between switches in the same fabric (intra-fabric links). PTP is created for inter-fabric links, and ttag is created for the inter-fabric link if the other fabric (Switch) is not managed by DCNM. Inter-fabric links do not support PTP or ttag configurations if both fabrics are managed by DCNM.
-
TTAG configuration is configured by default after the breakout. After the links are discovered and connected post breakout, perform Save & Deploy to generate the correct configuration based on the type of port (host, intra-fabric link, or inter fabric link).
Sync up Out-of-Band Switch Interface Configurations with DCNM
From DCNM release 11.5(1), any interface level configuration made outside of DCNM (via CLI) can be synced to DCNM and then managed from DCNM. Also, the vPC pair configurations are automatically detected and paired. This applies to the External_Fabric_11_1 and LAN_Classic fabrics only. The vPC pairing is performed with the vpc_pair policy.
Note |
When DCNM is managing switches, ensure that all configuration changes are initiated from DCNM and avoid making changes directly on the switch. |
When the interface config is synced up to the DCNM intent, the switch configs are considered as the reference, that is, at the end of the sync up, the DCNM intent reflects what is present on the switch. If there were any undeployed intent on DCNM for those interfaces before the resync operation, they will be lost.
Guidelines
-
Supported in fabrics using the following templates: Easy_Fabric_11_1, External_Fabric_11_1, and LAN_Classic.
-
Supported for Cisco Nexus switches only.
-
Supported for interfaces that don’t have any fabric underlay related policy associated with them prior to the resync. For example, IFC interfaces and intra fabric links aren’t subjected to resync.
-
Supported for interfaces that do not have any custom policy (policy template that isn’t shipped with Cisco DCNM) associated with them prior to resync.
-
Supported for interfaces where the intent is not exclusively owned by a Cisco DCNM feature and/or application prior to resync.
-
Supported on switches that don’t have Interface Groups associated with them.
-
Interface mode (switchport to routed, trunk to access, and so on) changes aren’t supported with overlays attached to that interface.
The sync up functionality is supported for the following interface modes and policies:
Interface Mode | Policies |
trunk (standalone, po, and vPC PO) |
|
access (standalone, po, and vPC PO) |
|
dot1q-tunnel |
|
routed |
int_routed_host_11_1 |
loopback |
int_freeform |
sub-interface |
int_subif_11_1 |
FEX (ST, AA) |
|
breakout |
interface_breakout |
nve | int_freeform (only in External_Fabric_11_1/LAN_Classic) |
SVI | int_freeform (only in External_Fabric_11_1/LAN_Classic) |
mgmt0 | int_mgmt_11_1 |
In an Easy fabric, the interface resync will automatically update the network overlay attachments based on the access VLAN or allowed VLANs on the interface.
After the resync operation is completed, the switch interface intent can be managed using normal DCNM procedures.
Syncing up Switch Interface Configurations to DCNM
Before you begin
-
We recommend taking a fabric backup before attempting the interface resync.
-
In External_Fabric_11_1 and LAN_Classic fabrics, for the vPC pairing to work correctly, both the switches must be in the fabric and must be functional.
-
Ensure that the switches are In-Sync and not in Migration-mode or Maintenance-mode.
Procedure
Step 1 |
In DCNM, navigate to Control > Fabric Builder and click a fabric. |
||||
Step 2 |
Ensure that switches are present in the fabric and vPC pairings are completed, and they are shown in the Topology view. Click Tabular view in the Actions panel. |
||||
Step 3 |
From Tabular view, select one or more switches where the interface intent resync is needed, and click Policies.
|
||||
Step 4 |
In the Policies window, click the Add Policy icon. |
||||
Step 5 |
In the Add Policy window, select host_port_resync from the Policy drop-down list. Click Save. |
||||
Step 6 |
Check the Mode column for the switches to ensure that they report Migration. For a vPC pair, both switches are in the Migration-mode.
|
||||
Step 7 |
After the configuration changes are ready to sync up to DCNM, navigate to the Tabular view, select the required switches, and click Rediscover switch to ensure that DCNM is aware of any new interfaces and other changes. |
||||
Step 8 |
Click Save & Deploy to start the resync process.
|
||||
Step 9 |
The Config Deployment window is displayed if no errors are detected during the resync operation. The interface intent is updated in DCNM.
Close the Config Deployment window, and you can see that the switches are automatically moved out of the Migration-mode. Switches in a vPC pair that were not paired or paired with no_policy show up as paired and associated with the vpc_pair policy.
|
What to do next
The following limitations are applicable after Syncing up Switch Interface Configurations to DCNM:
-
The port channel membership (once the policy exists) is not supported.
-
Changing the interface mode (trunk to access etc.) that have overlays attached is not supported.
-
Resync for interfaces that belong to Interface Groups are not supported.
-
The vPC pairing in External_Fabric_11_1 and LAN_Classic templates must be updated with the vpc_pair policy.
-
Changing the interface mode that have overlays attached is not supported.
-
In Easy_Fabric fabrics, VXLAN overlay interface attachments are performed automatically based on the allowed VLANs.
MACsec Support in Easy Fabric and eBGP Fabric
From Cisco DCNM Release 11.5(1), MACsec is supported in the Easy Fabric and eBGP Fabric on intra-fabric links. You should enable MACsec on the fabric and on each required intra-fabric link to configure MACsec. Unlike CloudSec, auto-configuration of MACsec is not supported.
MACsec is supported on switches with minimum Cisco NX-OS Releases 7.0(3)I7(8) and 9.3(5).
Note |
Support for MACsec is a preview feature in the Cisco DCNM Release 11.5(1). |
Guidelines
-
If MACsec cannot be configured on the physical interfaces of the link, an error is displayed when you click Save. MACsec cannot be configured on the device and link due to the following reasons:
-
The minimum NX-OS version is not met.
-
The interface is not MACsec capable.
-
-
MACsec global parameters in the fabric settings can be changed at any time.
-
MACsec and CloudSec can coexist on a BGW device.
-
MACsec is not supported on Border Leaf.
-
MACsec status of a link with MACsec enabled is displayed on the Links window.
-
Brownfield migration of devices with MACsec configured is supported using switch and interface freeform configs.
For more information about MACsec configuration, which includes supported platforms and releases, see the Configuring MACsec chapter in Cisco Nexus 9000 Series NX-OS Security Configuration Guide.
The following sections show how to enable and disable MACsec in DCNM:
Enabling MACsec
Procedure
Step 1 |
Navigate to Control > Fabrics > Fabric Builder. |
||||
Step 2 |
Click Create Fabric to create a new fabric or click Edit Fabric on an existing Easy or eBGP fabric. |
||||
Step 3 |
Click the Advanced tab and specify the MACsec details. Enable MACsec – Select the check box to enable MACsec for the fabric. MACsec Primary Key String – Specify a Cisco Type 7 encrypted octet string that is used for establishing the primary MACsec session. For AES_256_CMAC, the key string length must be 130 and for AES_128_CMAC, the key string length must be 66. If these values are not specified correctly, an error is displayed when you save the fabric.
MACsec Primary Cryptographic Algorithm – Choose the cryptographic algorithm used for the primary key string. It can be AES_128_CMAC or AES_256_CMAC. The default value is AES_128_CMAC. You can configure a fallback key on the device to initiate a backup session if the primary session fails. MACsec Fallback Key String - Specify a Cisco Type 7 encrypted octet string that is used for establishing a fallback MACsec session. For AES_256_CMAC, the key string length must be 130 and for AES_128_CMAC, the key string length must be 66. If these values are not specified correctly, an error is displayed when you save the fabric. MACsec Fallback Cryptographic Algorithm - Choose the cryptographic algorithm used for the fallback key string. It can be AES_128_CMAC or AES_256_CMAC. The default value is AES_128_CMAC. MACsec Cipher Suite – Choose one of the following MACsec cipher suites for the MACsec policy:
The default value is GCM-AES-XPN-256.
MACsec Status Report Timer - Specifies MACsec operational status periodic report timer in minutes. |
||||
Step 4 |
Click a fabric, click Tabular View in the Actions panel, and then click Links. |
||||
Step 5 |
Choose an intra-fabric link on which you want to enable MACsec and click Update Link. |
||||
Step 6 |
In the Link Management – Edit Link window, click Advanced in the Link Profile section, and select the Enable MACsec check box. If MACsec is enabled on the intra fabric link but not in the fabric settings, an error is displayed when you click Save. When MACsec is configured on the link, the following configurations are generated:
|
||||
Step 7 |
Click Save and then click Save & Deploy to deploy the MACsec configuration. |
Disabling MACsec
To disable MACsec on an intra-fabric link, navigate to the Link Management – Edit Link window, unselect the Enable MACsec check box, click Save, and then click Save & Deploy. This action performs the following:
-
Deletes MACsec interface policies from the link.
-
If this is the last link where MACsec is enabled, MACsec global policies are also deleted from the device.
Only after disabling MACsec on links, navigate to the Fabric Settings and unselect the Enable MACsec check box under the Advanced tab to disable MACsec on the fabric. If there’s an intra-fabric link in the fabric with MACsec enabled, an error is displayed when you click Save & Deploy.
Overview of Tenant Routed Multicast
Tenant Routed Multicast (TRM) enables multicast forwarding on the VXLAN fabric that uses a BGP-based EVPN control plane. TRM provides multi-tenancy aware multicast forwarding between senders and receivers within the same or different subnet local or across VTEPs.
With TRM enabled, multicast forwarding in the underlay is leveraged to replicate VXLAN encapsulated routed multicast traffic. A Default Multicast Distribution Tree (Default-MDT) is built per-VRF. This is an addition to the existing multicast groups for Layer-2 VNI Broadcast, Unknown Unicast, and Layer-2 multicast replication group. The individual multicast group addresses in the overlay are mapped to the respective underlay multicast address for replication and transport. The advantage of using a BGP-based approach allows the VXLAN BGP EVPN fabric with TRM to operate as fully distributed Overlay Rendezvous-Point (RP), with the RP presence on every edge-device (VTEP).
A multicast-enabled data center fabric is typically part of an overall multicast network. Multicast sources, receivers, and multicast rendezvous points might reside inside the data center but also might be inside the campus or externally reachable via the WAN. TRM allows a seamless integration with existing multicast networks. It can leverage multicast rendezvous points external to the fabric. Furthermore, TRM allows for tenant-aware external connectivity using Layer-3 physical interfaces or subinterfaces.
For more information, see the following:
Overview of Tenant Routed Multicast with VXLAN EVPN Multi-Site
Tenant Routed Multicast with Multi-Site enables multicast forwarding across multiple VXLAN EVPN fabrics connected via Multi-Site.
The following two use cases are supported:
-
Use Case 1: TRM provides Layer 2 and Layer 3 multicast services across sites for sources and receivers across different sites.
-
Use Case 2: Extending TRM functionality from VXLAN fabric to sources receivers external to the fabric.
TRM Multi-Site is an extension of BGP-based TRM solution that enables multiple TRM sites with multiple VTEPs to connect to each other to provide multicast services across sites in most efficient possible way. Each TRM site is operating independently and border gateway on each site allows stitching across each site. There can be multiple Border Gateways for each site. In a given site, the BGW peers with Route Sever or BGWs of other sites to exchange EVPN and MVPN routes. On the BGW, BGP will import routes into the local VRF/L3VNI/L2VNI and then advertise those imported routes into the Fabric or WAN depending on where the routes were learnt from.
Tenant Routed Multicast with VXLAN EVPN Multi-Site Operations
The operations for TRM with VXLAN EVPN Multi-Site are as follows:
-
Each Site is represented by Anycast VTEP BGWs. DF election across BGWs ensures no packet duplication.
-
Traffic between Border Gateways uses ingress replication mechanism. Traffic is encapsulated with VXLAN header followed by IP header.
-
Each Site will only receive one copy of the packet.
-
Multicast source and receiver information across sites is propagated by BGP protocol on the Border Gateways configured with TRM.
-
BGW on each site receives the multicast packet and re-encapsulate the packet before sending it to the local site.
For information about guidelines and limitations for TRM with VXLAN EVPN Multi-Site, see Configuring Tenant Routed Multicast.
Configuring TRM for Single Site Using Cisco DCNM
This section is assumes that a VXLAN EVPN fabric has already been provisioned using Cisco DCNM.
Procedure
Step 1 |
Enable TRM for the selected Easy Fabric. If the fabric template is Easy_Fabric_11_1, click the Fabric settings, navigate to the Replication tab, and check the Enable Tenant Routed Multicast (TRM) field. In addition, the default MDT multicast group field is auto-populated with a default value. Enable Tenant Routed Multicast (TRM): Select the check box to enable Tenant Routed Multicast (TRM) that allows overlay multicast traffic to be supported over EVPN/MVPN in the VXLAN BGP EVPN fabric. Default MDT Address for TRM VRFs: The multicast address for Tenant Routed Multicast traffic is populated. By default, this address is from the IP prefix specified in the Multicast Group Subnet field. When you update either field, ensure that the TRM address is chosen from the IP prefix specified in Multicast Group Subnet. Click Save to save the fabric settings. At this point, all the switches turn “Blue” as it will be in the pending state. Click Save and Deploy to enable the following:
For VXLAN EVPN fabric created using Easy_Fabric_eBGP fabric template, Enable Tenant Routed Multicast (TRM) field and Default MDT Address for TRM VRFs field can be found on the fabric settings' EVPN tab. |
||||
Step 2 |
Enable TRM for the VRF. Navigate to Control > VRFs and edit the selected VRF. Navigate to the Advanced Tab and edit the following TRM settings: TRM Enable – Select the check box to enable TRM. If you enable TRM, then the RP address and the underlay multicast address must be entered. Is RP External – Enable this check box if the RP is external to the fabric. If this field is unchecked, RP is distributed in every VTEP.
RP Address – Specifies the IP address of the RP. RP Loopback ID – Specifies the loopback ID of the RP, if Is RP External is not enabled. Underlay Multicast Address – Specifies the multicast address associated with the VRF. The multicast address is used for transporting multicast traffic in the fabric underlay.
Overlay Multicast Groups – Specifies the multicast group subnet for the specified RP. The value is the group range in “ip pim rp-address” command. If the field is empty, 224.0.0.0/24 is used as default. Click Save to save the settings. The switches go into the pending state, that is, blue color. These settings enable the following:
|
||||
Step 3 |
Enable TRM for the network. Navigate to Control > Networks. Edit the selected network and navigate to the Advanced tab. Edit the following TRM setting: TRM enable – Select the check box to enable TRM. Click Save to save the settings. The switches go into the pending state, that is, the blue color. The TRM settings enable the following:
|
Configuring TRM for Multi-Site Using Cisco DCNM
This section assumes that a Multi-Site Domain (MSD) has already been deployed by Cisco DCNM and TRM needs to be enabled.
Procedure
Step 1 |
Enable TRM on the BGWs. Navigate to Control > VRFs. Make sure that the right DC Fabric is selected under the Scope and edit the VRF. Navigate to the Advanced tab. Edit the TRM settings. Repeat this process for every DC Fabric and its VRFs. TRM Enable – Select the check box to enable TRM. If you enable TRM, then the RP address and the underlay multicast address must be entered. Is RP External – Enable this check box if the RP is external to the fabric. If this field is unchecked, RP is distributed in every VTEP.
RP Address – Specifies the IP address of the RP. RP Loopback ID – Specifies the loopback ID of the RP, if Is RP External is not enabled. Underlay Multicast Address – Specifies the multicast address associated with the VRF. The multicast address is used for transporting multicast traffic in the fabric underlay.
Overlay Multicast Groups – Specifies the multicast group subnet for the specified RP. The value is the group range in “ip pim rp-address” command. If the field is empty, 224.0.0.0/24 is used as default. Enable TRM BGW MSite - Select the check box to enable TRM on Border Gateway Multi-Site. Click on Save to save the settings. The switches go into the pending state, that is, blue color. These settings enable the following:
|
||||
Step 2 |
Establish MVPN AFI between the BGWs. Navigate to Control > Fabrics. Select the MSD fabric. Click Tabular view and click Links. Filter it by the policy - Overlays. Select and edit each overlay peering to enable TRM by checking the Enable TRM check box. Click Save to save the settings. The switches go into the pending state, that is, the blue color. The TRM settings enable the MVPN peering’s between the BGWs, or BGWs and Route Server. |
SSH Key RSA Handling
Bootstrap scenario
If the switch has the ssh key rsa command with the key-length variable value other than 1024 in the running configuration, the ssh key rsa key-length force command needs to be added to the bootstrap freeform configuration with the required value (any value other than 1024) during bootstrap.
Greenfield and Brownfield scenarios
Use the ssh key rsa key-length force command to change the key-length variable to a value other than 1024.
However, on Cisco Nexus 9000 Releases 9.3(1) and 9.3(2), the ssh key rsa key-length force command fails while the device is booting up during the ASCII replay process. For more information, refer CSCvs40704.
The configurations are considered to be in-sync when both the intent and switch running configurations have the same command. For example, the status is considered to be in-sync when the ssh key rsa 2048 command is present in both in the intent and the running configuration. However, consider a scenario in which the ssh key rsa 2040 command was pushed to the switch as an Out-Of-Band change. While the intent has a key-length value of 2048, the device has a key-length value of 2040. In such instances, the switch will be marked as out-of-sync.
The diff shown in the Pending Config tab (in both Strict Config-Compliance and non-Strict Config-Compliance mode) cannot be deployed onto the switch from DCNM as the feature ssh command has to be used to disable the SSH feature before making any change to the ssh key rsa command. This would lead to a dropped connection to DCNM. In such a scenario, the diff can be resolved by modifying the intent such that there is no diff.
With Strict Config-Compliance mode:
- Delete the Policy Template Instance (PTI) that has the ssh key rsa 2048 force command by clicking View/Edit Policies in the Tabular View of the Fabric Builder window.
- Create a new PTI with the ssh key rsa 2040 force command by clicking View/Edit Policies.
Without Strict Config-Compliance mode:
- Delete the PTI with the ssh key rsa 2048 force command in the intent by clicking View/Edit Policies in the Tabular View of the Fabric Builder window.
- Create a switch_freeform PTI with the ssh key rsa 2040 force command in the intent to match the Out-Of-Band change from the device.
Switch Operations
To view various options, right-click on switch:
Set Role: Assign a role to the switch. You can assign any one of the following roles to a switch:
-
Spine
-
Leaf (Default role)
-
Border
-
Border Spine
-
Border Gateway
-
Access
-
Aggregation
-
Edge Router
-
Core Router
-
Super Spine
-
Border Super Spine
-
Border Gateway Spine
-
ToR
Alternatively, you can also navigate to the Tabular view from the Actions pane. Choose one or more devices of the same device type and click Set Role to set roles for devices. The device types are:
-
NX-OS
-
IOS XE
-
IOS XR
-
Other
Note |
Ensure that you have moved switches from maintenance mode to active mode or operational mode before setting roles. You can change the switch role only before executing Save & Deploy. |
You can assign one of the following roles for non-Nexus devices:
-
Spine
-
Leaf
-
Access (This role is available only for Cisco ASR 1000 Series routers and Cisco Catalyst 9000 Series switches).
-
Edge Router (Use this role for VRF-Lite).
-
Core Router
-
Super Spine
-
Preview Config
-
ToR (This role is available only for Cisco Catalyst 9000 series switches).
From DCNM 11.1(1) release, you can shift the switch role from existing to required role if there are no overlays on the switches. Click Save and Deploy to generate the updated configuration. The following shifts are allowed for the switch role:
-
Leaf to Border
-
Border to Leaf
-
Leaf to Border Gateway
-
Border Gateway to Leaf
-
Border to Border Gateway
-
Border Gateway to Border
-
Spine to Border Spine
-
Border Spine to Spine
-
Spine to Border Gateway Spine
-
Border Gateway Spine to Spine
-
Border Spine to Border Gateway Spine
-
Border Gateway Spine to Border Spine
You cannot change the switch role from any Leaf role to any Spine role and from any Spine role to any Leaf role.
Switch[<serial-number>]: Role change from <switch-role> to <switch-role> is not permitted.
You can then change the switch role to the role that was set earlier, or set a new role, and configure the fabric.
If you have not created any policy template instances before clicking Save and Deploy, and there are no overlays, you can change the role of a switch to any other required role.
Switches role should be the same for VPC pairing. peer1 <serial-number>: [<switch-role>], peer2 <serial-number>: [<switch-role>]
To prevent this scenario, change the switch roles of both the switches in the vPC pair to the same role.
Running EXEC Mode Commands in DCNM
When you first log in, the Cisco NX-OS software places you in the EXEC mode. The commands available in the EXEC mode include the show commands that display the device status and configuration information, the clear commands, and other commands that perform actions that you do not save in the device configuration.
The following procedure shows how to run EXEC commands in DCNM:
Procedure
Step 1 |
From DCNM, navigate to Control > Fabrics > Fabric Builder. |
Step 2 |
Click a fabric and then click Tabular view in the Actions menu. |
Step 3 |
Select one or more switches and click the Play button (Execute Commands). |
Step 4 |
From the Template drop-down list, select exec_freeform. |
Step 5 |
Enter the commands in the Freeform CLI field. |
Step 6 |
Click Deploy to run the EXEC commands. |
Step 7 |
In the CLI Execution Status window, you can check the status of the deployment. Click Detailed Status under the Command column to view details. |
Step 8 |
In the Command Execution Details window, click the info under the CLI Response column to view the output or response. |
Fabric Multi Switch Operations
Click Tabular view from the Actions pane in the fabric topology window. The tabular view has the following tabs:
Tabular View - Switches
You can manage switch operations in this tab. Each row represents a switch in the fabric, and displays switch details, including its serial number.
Some of the actions that you can perform from this tab are also available when you right-click a switch in the fabric topology window. However, the Switches tab enables you to provision configurations on multiple switches, like deploying policies, simultaneously.
The Switches tab has following information of every switch you discover in the fabric:
-
Name: Specifies the switch name.
-
IP Address: Specifies the IP address of the switch.
-
Role: Specifies the role of the switch.
-
Serial Number: Specifies the serial number of the switch.
-
Fabric Name: Specifies the name of the fabric, where the switch is discovered.
-
Fabric Status: Specifies the status of the fabric, where the switch is discovered.
-
Discover Status: Specifies the discovery status of the switch.
-
Model: Specifies the switch model.
-
Software Version: Specifies the software version of the switch.
-
ThousandEyes Status: Specifies the status of the ThousandEyes Enterprise Agent.
-
Last Updated: Specifies when the switch was last updated.
-
Mode: Specifies the current mode of the switch.
-
VPC Role: Specifies the vPC role of the switch.
-
VPC Peer: Specifies the vPC peer of the switch.
The Switches tab has the following icons and buttons:
-
Add switches: Click this icon to discover existing or new switches to the fabric. The Inventory Management dialog box appears.
This option is also available in the fabric topology window. Click Add switches in the Actions pane.
Refer the following sections for more information:
-
Adding Switches to a Fabric: Provides information on adding switches to easy fabrics.
-
Discovering New Switches: Provide information on adding Cisco Nexus switches to external fabrics.
-
Adding non-Nexus Devices to External Fabrics: Provide information on adding non-Nexus switches to external fabrics.
-
-
Rediscover switch: Initiate the switch discovery process by DCNM afresh.
-
Update discovery credentials: Update device credentials such as authentication protocol, username and password.
-
Saving config and Reload: Save the configurations and reload the switch.
Note
This option is grayed out if the fabric is in freeze mode, that is, if you have disabled deployments on the fabric.
-
Copy running to startup config: From Cisco DCNM, Release 11.4(1), you can perform an on-demand copy running-configuration to startup-configuration operation for one or more switches.
Note
This option will be grayed out if the fabric is in freeze mode, that is, if you have disabled deployments on the fabric.
-
Remove switches: Remove the switch from the fabric.
Note
This option will be grayed out if the fabric is in freeze mode, that is, if you have disabled deployments on the fabric.
-
Preview: You can preview the pending configurations and the side-by-side comparison of running configurations and expected configurations.
-
Policies: Add, update and delete a policy. The policies are template instances of templates in the template library. After creating a policy, you should deploy it on the switches using the Deploy option available in the Policies window. You can select more than one policy and view them.
Note
If you select multiple switches and deploy a policy instance, then it will be deployed on all the selected switches.
-
ThousandEyes Agent: You can start, stop, install, or uninstall ThousandEyes Enterprise Agent on the switch. You can choose single or multiple switches and select required operation from ThousandEyes Agent drop-down list.
Note
When you choose multiple switches to perform ThousandEyes Enterprise Agent action, ensure that the status of selected switches are same.
-
Interfaces: Deploy configurations on the switch interfaces.
-
History: View the deployment history and the policy change history using this button. Choose one or more switches and click History.
The Policy Change History tab lists the history of policies along with the users who made the changes like add, update, or delete.
Under the Policy Change History tab, for a policy, click Detailed History under the Generated Config column to view the generated config before and after.
The following table provides the summary of generated config before and after for Policy Template Instances (PTIs).
PTI Operations Generated Config Before Generated Config After Add Empty Contains the config Update Contains config before changes Contains config after changes Mark-Delete Contains the config to be removed Contains the config to be removed with colour change Delete Contains the config Empty Note
When a policy or profile template is applied, an instance is created for each application of the template. This instance is known as Policy Template Instance or PTI.
-
Deploy: Deploy switch configurations. From Cisco DCNM Release 11.3(1), you can deploy configurations for multiple devices using the Deploy button.
Note
-
This option grays out if the fabric is in freeze mode, that is, if you have disabled deployments on the fabric.
-
In an MSD fabric, you can deploy configurations only on the Border Gateway, Border Gateway Spine, Border Gateway Super-Spine, or External Fabric switches.
-
-
Set Role: Choose one or more devices of the same device type and click Set Role to set roles for devices. The device types are:
-
NX-OS
-
IOS XE
-
IOS XR
-
Other
Ensure that you have moved switches from maintenance mode to active mode or operational mode before setting roles. See the Switch Operations section for more information on setting roles.
-
-
vPC Pairing: Choose a switch and click vPC Pairing to create, edit, or unpair a vPC pair. However, you can use this option only when you choose a Cisco Nexus switch. Refer the following sections for more information:
-
Creating a vPC Setup in the External Fabric: Provides information on how to create a vPC pair in external fabrics.
-
vPC Fabric Peering: Provides information on how to create a vPC pair in easy fabrics.
-
Tabular View - Links
You can add links between border switches of different fabrics (inter-fabric links) or between switches in the same fabric (intra-fabric links). You can only create an inter-fabric connection (IFC) for a switch that is managed by DCNM.
There are scenarios where you might want to define links between switches before connecting them physically. The links could be inter-fabric or intra-fabric links. Doing so, you can express and represent your intent to add links. The links with intent are displayed in a different colour till they are actually converted to functional links. Once you physically connect the links, they are displayed as connected.
Management links might show up in the fabric topology as red colored links. To remove such links, right-click the link and click Delete Link.
From Cisco DCNM Release 11.1(1), the Border Spine and Border Gateway Spine roles are added to switch roles for border switches.
You can create links between existing and pre-provisioned devices as well by selecting the pre-provisioned device as the destination device.
Creating Intra-Fabric Links
-
Click Control > Fabric Builder to go to the Fabric Builder screen.
-
Click within the rectangular box that represents the fabric. The fabric topology screen comes up.
-
Click Tabular view in the Actions panel that is displayed at the left part of the screen.
A screen with the tabs Switches and Links appears. They list the fabric switches and links in a table.
-
Click the Links tab. You can see a list of links.
The list is empty when you are yet to create a link.
-
Click the Add (+) button at the top left part of the screen to add a link.
The Add Link screen comes up. By default, the Intra-Fabric option is chosen as the link type.
The fields are:
Link Type – Choose Intra-Fabric to create a link between two switches in a fabric.
Link Sub-Type – This field populates Fabric indicating that this is a link within the fabric.
Link Template: You can choose any of the following link templates.
-
int_intra_fabric_num_link_11_1: If the link is between two ethernet interfaces assigned with IP addresses, choose int_intra_fabric_num_link_11_1.
-
int_intra_fabric_unnum_link_11_1: If the link is between two IP unnumbered interfaces, choose int_intra_fabric_unnum_link_11_1.
-
int_intra_vpc_peer_keep_alive_link_11_1: If the link is a vPC peer keep-alive link, choose int_intra_vpc_peer_keep_alive_link_11_1.
-
int_pre_provision_intra_fabric_link: If the link is between two pre-provisioned devices, choose int_pre_provision_intra_fabric_link. After you click Save & Deploy, an IP address is picked from the underlay subnet IP pool.
Correspondingly, the Link Profile section fields is updated.
Source Fabric – The fabric name populates this field since the source fabric is known.
Destination Fabric – Choose the destination fabric. For an intra-fabric link, source and destination fabrics are the same.
Source Device and Source Interface – Choose the source device and interface.
Destination Device and Destination Interface – Choose the destination device and interface.
Note
Select the pre-provisioned device as the destination device if you are creating a link between an existing device and a pre-provisioned device.
General tab in the Link Profile section
Interface VRF – Name of a non-default VRF for this interface.
Source IP and Destination IP – Specify the source and destination IP addresses of the source and destination interfaces, respectively.
Note
The Source IP and Destination IP fields do not appear if you choose int_pre_provision_intra_fabric_link template.
Interface Admin State – Check or uncheck the check box to enable or disable the admin sate of the interface.
MTU – Specify the maximum transmission unit (MTU) through the two interfaces.
Advanced tab.
Source Interface Description and Destination Interface Description – Describe the links for later use. For example, if the link is between a leaf switch and a route reflector device, you can enter the information in these fields (Link from leaf switch to RR 1 and Link from RR 1 to leaf switch). This description will converted into a config, but will not be pushed into the switch. After Save & Deploy, it will reflect in the running configuration.
Disable BFD Echo on Source Interface and Disable BFD Echo on Destination Interface – Select the check box to disable BFD echo packets on source and destination interface.
Note that the BFD echo fields are applicable only when you have enabled BFD in the fabric settings.
Source Interface Freeform CLIs and Destination Interface Freeform CLIs: Enter the freeform configurations specific to the source and destination interfaces. You should add the configurations as displayed in the running configuration of the switch, without indentation. For more information, refer Enabling Freeform Configurations on Fabric Switches.
-
-
Click Save at the bottom right part of the screen.
The new link appears in the Links tab.
-
Click Save & Deploy to deploy the link configurations on the switches.
The Config Deployment screen comes up. It displays the configuration status on the switches. You can also view the pending configurations by clicking the respective link in the Preview Config column. When you click a link in the Preview Config column, the Config Preview window comes up. It lists the pending configurations on the switch. The Side-by-side Comparison tab displays the running configuration and expected configuration side-by-side.
-
Close the preview screen and click Deploy Config. The pending configurations are deployed.
-
After ensuring that the progress is 100% in all the rows, click Close at the bottom part of the screen. The Links screen comes up again.
Click <- at the top left part of the screen to go to the fabric topology. In the fabric topology, you can see that the link between the two devices is displayed.
Creating Inter-Fabric Links
-
Click the Links tab in the Switches | Links page. The list of previously created links is displayed. The list contains intra-fabric links (between switches in a fabric), and inter-fabric links (between BGWs or border leaf/spine switches of different fabrics).
-
Click the Add (+) button at the top left part of the screen to add a link. The Add Link screen comes up.
By default, the Intra-Fabric option is chosen as the link type.
-
From the Link Type drop-down box, choose Inter-Fabric since you are creating an IFC. The screen changes correspondingly.
The fields for inter-fabric link creation are explained:
Link Type – Choose Inter-Fabric to create an inter-fabric connection between two fabrics, via their border switches.
Link Sub-Type – This field populates the IFC type. Choose VRF_LITE, MULTISITE_UNDERLAY, or MULTISITE_OVERLAY from the drop-down list.
The Multi-Site options are explained in the Multi-Site use case.
For information about VXLAN MPLS interconnection, see the Border Provisioning Use Case in VXLAN BGP EVPN Fabrics - MPLS SR and LDP Handoff chapter.
For information about routed fabric interconnection, see the Creating Inter-Fabric Links Between a Routed Fabric and an External Fabric section in the Configuring a Fabric with eBGP Underlay chapter.
Link Template: The link template is populated.
The templates are autopopulated with corresponding pre-packaged default templates that are based on your selection.
Note
You can add, edit, or delete user-defined templates. See Template Library section in the Control chapter for more details.
Source Fabric - This field is prepopulated with the source fabric name.
Destination Fabric - Choose the destination fabric from this drop-down box.
Source Device and Source Interface - Choose the source device and Ethernet interface that connects to the destination device.
Destination Device and Destination Interface—Choose the destination device and Ethernet interface that connects to the source device.
Based on the selection of the source device and source interface, the destination information is autopopulated based on Cisco Discovery Protocol information, if available. There is an extra validation performed to ensure that the destination external device is indeed part of the destination fabric.
General tab in the Link Profile section.
Local BGP AS# - In this field, the AS number of the source fabric is autopopulated.
IP_MASK—Fill up this field with the IP address of the source interface that connects to the destination device.
NEIGHBOR_IP—Fill up this field with the IP address of the destination interface.
NEIGHBOR_ASN—In this field, the AS number of the destination device is autopopulated.
After filling up the Add Link screen, it looks like this:
-
Click Save at the bottom right part of the screen.
The Switches|Links screen comes up again. You can see that the IFC is created and displayed in the list of links.
-
Click on Save & Deploy to deploy the link configurations on the switches.
The Config Deployment screen comes up. It displays the configuration status on the switches. You can also view the pending configurations by clicking the respective link in the Preview Config column. When you click a link in the Preview Config column, the Config Preview window comes up. It lists the pending configurations on the switch. The Side-by-side Comparison tab displays the running configuration and expected configuration side-by-side.
-
Close the preview screen and click Deploy Config. The pending configurations are deployed.
-
After ensuring that the progress is 100% in all the rows, click Close at the bottom part of the screen. The Links screen comes up again.
-
Click <- at the top left part of the screen to go to the fabric topology. In the fabric topology, you can see that the link between the two devices is displayed.
If the two fabrics are member fabric of an MSD, then you can see the link in the MSD topology too.
When you enable the VRF Lite function using the ToExternalOnly method or Multisite function via MSD fabric, IFCs are automatically created between the (VXLAN fabric) border/BGW device and connected (external fabric) edge router/core device. When you remove the ER/core/border/BGW device, the corresponding IFCs (link PTIs) to/from that switch are deleted on DCNM. Subsequently, DCNM removes the corresponding IFC configurations, if any, from the remaining devices on the next Save & Deploy operation. Also, if you want to remove a device that has an IFCs and overlay extensions over those IFCs, you should undeploy all overlay extensions corresponding to those IFCs for switch delete to be possible.
To undeploy VRF extensions, click Control > Networks & VRFs, select the VXLAN fabric and the extended VRFs, and undeploy the VRFs in the VRF deployment screen.
To delete the IFCs, click Control > Fabric Builder, go to the fabric topology screen, click Tabular view, and delete the IFCs from the Links tab.
Ensure that the fabric switch names are unique. If you deploy VRF extensions on switches with the same name, it leads to erroneous configuration.
The new fabric is created, the fabric switches are discovered in DCNM, the underlay networks provisioned on those switches, and the configurations between DCNM and the switches are synced. The remaining tasks are:
-
Provision interface configurations such as vPCs, loopback interface, and subinterface configurations. Refer Interfaces.
-
Create overlay networks and VRFs and deploy them on the switches. Refer Creating and Deploying Networks and VRFs.
Exporting Links
-
Choose
, and select a fabric.The fabric topology window appears.
-
Click Tabular view in the Actions panel.
A window with the Switches and Links tabs appears.
-
Click the Links tab.
You can see a list of links. The list is empty when you are yet to create a link.
-
Click the Export Links icon to export the links in a CSV file.
The following details of links are exported: link template, source fabric, destination fabric, source device, destination device, source switch name, destination switch name, source interface, destination interface, and nvPairs. The nvPairs field consists JSON object.
Importing Links
You can import a CSV file containing details of links to add new links to the fabric. The CSV file should have the following details of links: link template, source fabric, destination fabric, source device, destination device, source switch name, destination switch name, source interface, destination interface, and nvPairs.
Note |
|
-
Choose
, and select a fabric.The fabric topology window appears.
-
Click Tabular view in the Actions panel.
A window with the Switches and Links tabs appears.
-
Click the Links tab.
You can see a list of links. The list is empty when you are yet to create a link.
-
Click the Import Links icon.
The file server directory opens.
-
Browse the directory and select the CSV file that you want to import.
-
Click Open.
A confirmation screen appears.
-
Click Yes to import the selected file.
Viewing Details of Fabric Links
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder and select a fabric. The topology view of the fabric appears. |
||
Step 2 |
Double-click any of the links. The details window appears. You can view the devices that are connected using this link, summary, and the data traffic. |
||
Step 3 |
Click Show more details. A comparison table of the two devices connected by the link appears. It includes the following parameters of the devices: device name, name, admin status, operation status, reason, policies, overlay network, status, PC, vPC ID, speed, MTU, mode, VLANs, IP or prefix, VRF, neighbor, and description.
|
||
Step 4 |
Click the Back icon to go back to the details window.
|
Viewing the Traffic Details of Fabric Links
In the details window of a fabric link, you can choose how you want to view the traffic details. You can view the traffic details based on the time duration, format, and export this information.
You can view the data traffic of a link for the following durations from the duration drop-down list:
-
24 Hours
-
Week
-
Month
-
Year
Show: Click Show, and choose Chart, Table, or Chart and Table from the drop-down list to see how you want to view the traffic details. Enlarge your browser window to view the details in Chart and Table format.
If you choose Chart, hover over the traffic chart to view the Rx and Tx values, along the Y axis, for the corresponding time, along X axis. You can change the time duration values of the X axis by moving the sliders in the time range selector. You can choose the Y-axis values by checking or unchecking the Rx and Tx check boxes.
Note |
If you select Week, Month, or Year as the time duration, you can also view the Peak Rx and Peak Tx values along the Y axis. |
Select Table to view the traffic information in tabular format.
Chart Type and Chart Options: Choose Area Chart or Line Chart from the Chart Type drop-down list.
You can choose the following chart options:
-
Show Fill Patterns
-
Show Datamarkers
-
Y Axis Log Scale
Actions: Export or print the traffic information by choosing the appropriate options from the Actions drop-down list.
Symmetric Automatic VRF Lite
-
Check the Auto Deploy Flag check box in the Link Management dialog box. Checking this check box enables VRF lite deployment on both ends of the link for managed devices.
-
When you extend the VRF lite in a back-to-back scenario, the VRF should already be present in the peer fabric and the VRF name should be the same. An error message appears if the VRF is not present in the peer fabric and if you try to extend the VRF lite.
-
When you extend the VRF lite between an easy fabric and an external fabric, the VRF name can be the same as that of the source fabric, default, or another VRF name. However, the child PTIs for the subinterface and the VRF creation or peering on the external fabric has the source. Hence, you cannot edit or delete the policies from the View/Edit policies window.
-
If you perform a DCNM upgrade and notice that the policies are not attached to the IFC, edit the policies and VRF to attach them again.
-
Besides the IPv6 address, enter the IP mask, IPv4 address, and the neighbor IP address as well to deploy VRF from topdown using symmetric VRF lite.
-
Deploy configurations in both the fabrics.
-
You can edit or delete IFCs in the Link tab in the VXLAN fabric. The extra consideration for auto configured IFCs is that, in order to prevent the regeneration of IFC on next save and deploy, change the mode back to manual mode, or save the configuration only on the relevant devices.
-
In a back-to-back scenario, if you delete the VRF lite IFC on one of the fabrics, the VRF lite is deleted from the peer fabric as well.
-
When you want to delete a VRF lite between an easy fabric and an external fabric, delete the extension in the easy fabric using the top-down approach. The extension is automatically deleted from the external fabric.
-
Deploy the configurations in both the fabrics.
-
You must add redistribute hmm command in the freeform configuration when vrf-lite is configured on Border device.
See the Border Provisioning Use Case in VXLAN BGP EVPN Fabrics - VRF Lite chapter for a use case on VRF Lite.
Layer 3 Port Channels
From Cisco DCNM Release 11.3(1), Layer 3 port channels are supported in external links and interfaces. In the Interfaces window, you can select a port channel and a corresponding Layer 3 port channel interface template. This template allows you to configure various options related to Layer 3 port channels including an ability to specify all Layer 3 interface-related configurations. Layer 3 port channels are supported only in easy fabrics and external fabrics.
External connectivity using VRF_LITE will also be supported using Layer 3 port-channels. For physical routed interfaces and LAYER 3 port channel interfaces, you can set the MTU.
You can also watch the video that demonstrates how to extend symmetric VRF Lite using Layer 3 port channels in Cisco DCNM. See the Extending Symmetric VRF Lite Using Layer 3 Port Channels video.
Configuring Layer 3 Port Channel on Interfaces
To configure a Layer 3 port channel on an interface from the Cisco DCNM Web UI, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabrics > Interfaces. The Interfaces window appears. |
Step 2 |
Click Add Interface. The Add Interface dialog box appears. |
Step 3 |
Choose the Port Channel type and a device. The port-channel ID is autopopulated. |
Step 4 |
Choose the int_l3_port_channel policy. The fields under the General area changes accordingly. |
Step 5 |
Enter the values in the fields and click Save. Only saved configurations are pushed to the device. While adding the interface, you can only modify the policy attribute after the first save. If you try to use an ID that is already used, the Resource could not be allocated error appears. |
Step 6 |
(Optional) Click the Preview option to preview the configurations to be deployed. |
Step 7 |
Click Deploy to deploy the specified logical interface. The newly added interface appears in the screen. You can break out and unbreakout an interface by using the breakout option at the top left. |
Configuring Layer 3 Port Channel on Interfaces for IOS XE Devices
To configure a Layer 3 port channel on an interface for IOS XE devices, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabrics > Interfaces. The Interfaces window appears. |
||
Step 2 |
Click Add Interface. The Add Interface dialog box appears. |
||
Step 3 |
Choose the Port Channel type and a device. The port-channel ID is autopopulated. |
||
Step 4 |
Choose the ios_xe_int_l3_port_channel policy. The fields under the General area changes accordingly. |
||
Step 5 |
Enter the values in the fields and click Save. Only saved configurations are pushed to the device. While adding the interface, you can only modify the policy attribute after the first save. If you try to use an ID that is already used, the Resource could not be allocated error appears.
|
||
Step 6 |
(Optional) Click the Preview option to preview the configurations to be deployed. |
||
Step 7 |
Click Deploy to deploy the specified logical interface. The newly added interface appears in the screen. |
Deploying Policies on Physical Interfaces for non-Nexus Devices
More policies are added to support non-Nexus devices from Cisco DCNM, Release 11.4(1). When you import any non-Nexus device into an external fabric, few physical interfaces are created by default based on the number of ports. The policy is created only for the management port. For the Cisco Catalyst 9000 Series switches, the management port is GigabitEthernet0/0, and for the Cisco ASR 1000 Series routers, the management port is GigabitEthernet0.
The following table lists the policies added for different non-Nexus devices:
Devices |
Policies |
||
---|---|---|---|
Cisco CSR 1000V Series Router |
GigabitEthernet |
||
Cisco IOS-XE Devices |
|
To deploy policies on physical interfaces in the Interfaces window of Cisco DCNM Web UI, perform the following steps:
Before you begin
Import and discover non-Nexus devices into the external fabric. Ensure that the fabric isn’t in monitor mode.
Procedure
Step 1 |
Check the check box of the interface on which you want to deploy the policy. |
||
Step 2 |
Click the Edit Configuration icon. |
||
Step 3 |
Choose a policy from the Policy drop-down list. The valid options are:
|
||
Step 4 |
Enter values for all the mandatory fields.
|
||
Step 5 |
Click Save. |
||
Step 6 |
Click Preview to preview the pending configurations. |
||
Step 7 |
Click Deploy to deploy the policy on the interface. |
Configuring Layer 3 Port Channel on Subinterfaces
To configure a Layer 3 port channel on an interface from the Cisco DCNM Web UI, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabrics > Interfaces. The Interfaces window appears. |
Step 2 |
Choose a Layer 3 port channel interface. |
Step 3 |
Click Add Interface. The Add Interface dialog box appears. |
Step 4 |
Choose the Subinterface type. The subinterface ID and policy are autopopulated, and the fields under the General area changes accordingly. |
Step 5 |
Enter the values in the fields and click Save. Only saved configurations are pushed to the device. |
Step 6 |
(Optional) Click the Preview option to preview the configurations to be deployed. |
Step 7 |
Click Deploy to deploy the specified logical interface. A confirmation window appears, and the newly added subinterface appears in the list. |
Configuring Layer 3 Port Channel for Inter-fabric Connectivity
To configure a Layer 3 port channel link from the Fabric Builder window, perform the following steps:
Before you begin
Procedure
Step 1 |
Choose an easy fabric or an external fabric, for which you want to extend the VRF-Lite. The fabric topology window appears. |
Step 2 |
Click Tabular view in the Actions pane. All the components of this fabric are listed with their status and other details accordingly in different tabs. |
Step 3 |
Choose the Links tab. |
Step 4 |
Click the Add Link icon. The Add Link dialog box appears. |
Step 5 |
Choose Inter-Fabric link type. |
Step 6 |
Choose VRF_LITE link sub-type. |
Step 7 |
Choose the link template from the Link Template drop-down list. Valid values are ext_fabric_setup_11_1 and service_link_trunk. |
Step 8 |
Enter the details for all other fields accordingly. |
Step 9 |
Enter the details for fields in the Link Profile area wherever necessary. You can set the MTU. The Ext_VRF_Lite_Jython auto-deploy template is used for VRF-Lite configuration on the device in the fabric. |
Step 10 |
Click Save. |
What to do next
Tabular View - Operational View
From Cisco DCNM 11.3(1), the operational support for a fabric is provided. This feature provides the following information:
-
Operational status of a fabric
-
Alarm and event notifications
You can view the operational status information in the Operational View tab. You can view the alarm and event notifications by clicking the Alerts and Notifications icon, next to the Help icon, in the top pane of Cisco DCNM.
Viewing the Operational Status
To view the operational status of a fabric from the Fabric Builder window, perform the following steps:
Procedure
Step 1 |
Choose a fabric. |
||||||||||||||
Step 2 |
Click Tabular view in the Actions pane. |
||||||||||||||
Step 3 |
Choose the Operational View tab. The Operational View tab has the following fields and descriptions.
All these columns are sortable. |
Viewing Logical Links
The logical links appear in the Topology window. To view the logical links from the Cisco DCNM Web UI, perform the following steps:
Procedure
Step 1 |
Choose Topology. |
||
Step 2 |
Check the Logical Links check box in the Show pane. The logical links between devices appear in blue color.
|
||
Step 3 |
(Optional) Hover over the link to know the link type. |
Viewing Alerts and Event Notifications
Alert and event notifications includes health score, topology node display, alarm view, alarm policies, and notification services. An event is any action that impacts network, devices or Cisco DCNM. An alert is a notification that is triggered as part of an event to make it visible.
Support for ToR Switches
From Cisco DCNM 11.3(1), support for the Top-of-Rack (ToR) switches is added in DCNM. You can add the Layer 2 ToR switches in an external fabric, and they can be connected to the Leaf switches in the Easy Fabric. For more information, see Configuring ToR Switches and Deploying Networks.
vPC Fabric Peering
You can create a virtual peer link for two switches or change the existing physical peer link to a virtual peer link. Only greenfield deployments support vPC fabric peering in Cisco DCNM, Release 11.2(1). However, both greenfield as well as brownfield deployments support vPC fabric peering in Cisco DCNM, Release 11.3(1). This feature is applicable for Easy_Fabric_11_1 and Easy_Fabric_eBGP fabric templates.
Note |
The Easy_Fabric_eBGP fabric does not support brownfield import. |
Guidelines and Limitations
The following are the guidelines and limitations for vPC fabric pairing.
-
vPC fabric peering is supported from Cisco DCNM Release 11.2(1) and Cisco NX-OS Release 9.2(3).
-
Only Cisco Nexus N9K-C9332C Switch, Cisco Nexus N9K-C9364C Switch, Cisco Nexus N9K-C9348GC-FXP Switch as also the Cisco Nexus 9000 Series Switches that ends with FX, and FX2 support vPC fabric peering.
-
From Cisco DCNM, Release 11.4(1), Cisco Nexus N9K-C93180YC-FX3S and N9K-C93108TC-FX3P platform switches support vPC fabric peering.
-
Cisco Nexus 9300-EX, and 9300-FX/FXP/FX2/FX3 platform switches support vPC Fabric Peering. Cisco Nexus 9200 and 9500 platform switches do not support vPC Fabric Peering.
-
If you use other Cisco Nexus 9000 Series Switches, a warning will appear during Save & Deploy. A warning appears in this case because these switches will be supported in future releases.
-
If you try pairing switches that do not support vPC fabric peering, using the Use Virtual Peerlink option, a warning will appear when you deploy the fabric.
-
You can convert a physical peer link to a virtual peer link and vice-versa with or without overlays.
-
Switches with border gateway leaf roles do not support vPC fabric peering.
-
vPC fabric peering is not supported for Cisco Nexus 9000 Series Modular Chassis and FEXs. An error appears during Save & Deploy if you try to pair any of these.
-
Brownfield deployments and greenfield deployments support vPC fabric peering in Cisco DCNM, Release 11.3(1).
-
However, you can import switches that are connected using physical peer links and convert the physical peer links to virtual peer links after Save & Deploy. To update a TCAM region during the feature configuration, use the hardware access-list tcam ingress-flow redirect 512 command in the configuration terminal.
QoS for Fabric vPC-Peering
From Cisco DCNM Release 11.4(1), in the Easy_Fabric_11_1 fabric settings, you can enable QoS on spines for guaranteed delivery of vPC Fabric Peering communication. Additionally, you can specify the QoS policy name.
Note the following guidelines for a greenfield deployment:
-
If QoS is enabled and the fabric is newly created:
-
If spines or super spines neighbor is a virtual vPC, make sure neighbor is not honored from invalid links, for example, super spine to leaf or borders to spine when super spine is present.
-
Based on the Cisco Nexus 9000 Series Switch model, create the recommended global QoS config using the switch_freeform policy template.
-
Enable QoS on fabric links from spine to the correct neighbor.
-
-
If the QoS policy name is edited, make sure policy name change is honored everywhere, that is, global and links.
-
If QoS is disabled, delete all configuration related to QoS fabric vPC peering.
-
If there is no change, then honor the existing PTI.
For more information about a greenfield deployment, see the Creating a New VXLAN BGP EVPN Fabric section.
Note the following guidelines for a brownfield deployment:
Brownfield Scenario 1:
-
If QoS is enabled and the policy name is specified:
Note
You need to enable only when the policy name for the global QoS and neighbor link service policy is same for all the fabric vPC peering connected spines.
-
Capture the QoS config from switch based on the policy name and filter it from unaccounted configuration based on the policy name and put the configuration in the switch_freeform with PTI description.
-
Create service policy configuration for the fabric interfaces as well.
-
Greenfield config should make sure to honor the brownfield config.
-
-
If the QoS policy name is edited, delete the existing policies and brownfield extra configuration as well, and follow the greenfield flow with the recommended config.
-
If QoS is disabled, delete all the configuration related to QoS fabric vPC peering.
Note
No cross check for possible or error mismatch user configuration, and user might see the diff.
Brownfield Scenario 2:
-
If QoS is enabled and the policy name is not specified, QoS configuration is part of the unaccounted switch freeform config.
-
If QoS is enabled from fabric settings after Save & Deploy for brownfield, QoS configuration overlaps and you will see the diff if fabric vPC peering config is already present.
For more information about a brownfield deployment, see the Creating a New VXLAN BGP EVPN Fabric section..
Fields and Description
To view the vPC pairing window of a switch, from the fabric topology window, right-click the switch and choose vPC Pairing. The vPC pairing window for a switch has the following fields:
Field |
Description |
||
---|---|---|---|
Use Virtual Peerlink |
Allows you to enable or disable the virtual peer linking between switches. |
||
Switch name |
Specifies all the peer switches in a fabric.
|
||
Recommended |
Specifies if the peer switch can be paired with the selected switch. Valid values are true and false. Recommended peer switches will be set to true. |
||
Reason |
Specifies why the vPC pairing between the selected switch and the peer switches is possible or not possible. |
||
Serial Number |
Specifies the serial number of the peer switches. |
You can perform the following with the vPC Pairing option:
Creating a Virtual Peer Link
Procedure
Step 1 |
Choose Control > Fabrics. The Fabric Builder window appears. |
||
Step 2 |
Choose a fabric with the Easy_Fabric_11_1 or Easy_Fabric_eBGP fabric templates. The fabric topology window appears. |
||
Step 3 |
Right-click a switch and choose vPC Pairing from the drop-down list. The window to choose the peer appears.
You will get the following error when you choose a switch with the border gateway leaf role. <switch-name> has a Network/VRF
attached. Please detach the Network/VRF before vPC
Pairing/Unpairing |
||
Step 4 |
Check the Use Virtual Peerlink check box. |
||
Step 5 |
Choose a peer switch and check the Recommended column to see if pairing is possible. If the value is true, pairing is possible. You can pair switches even if the recommendation is false. However, you will get a warning or error during Save & Deploy. |
||
Step 6 |
Click Save. |
||
Step 7 |
In the Fabric Topology window, click Save & Deploy. The Config Deployment window appears. |
||
Step 8 |
Click the field against the switch in the Preview Config column. The Config Preview window appears for the switch. |
||
Step 9 |
View the vPC link details in the pending configuration and the side-by-side configuration. |
||
Step 10 |
Close the window. |
||
Step 11 |
Click the pending errors icon next to the Save & Deploy icon to view errors and warnings, if any. If you see any warnings that are related to TCAM, click the Resolve icon. A confirmation dialog box about reloading switches appears. Click OK. You can also reload the switches from Tabular view in the fabric topology window. The switches that are connected through vPC fabric peering, are enclosed in a gray cloud. |
Converting a Physical Peer Link to a Virtual Peer Link
Before you begin
-
Plan the conversion from physical peer link to virtual peer link during the maintenance window of switches.
-
Ensure the switches support vPC fabric peering. Only the following switches support vPC fabric peering:
-
Cisco Nexus N9K-C9332C Switch, Cisco Nexus N9K-C9364C Switch, and Cisco Nexus N9K-C9348GC-FXP Switch
-
Cisco Nexus 9000 Series Switches that ends with FX, FX2, and FX2-Z
-
Procedure
Step 1 |
Choose Control > Fabrics. The Fabric Builder window appears. |
||
Step 2 |
Choose a fabric with the Easy_Fabric_11_1 or Easy_Fabric_eBGP fabric templates. |
||
Step 3 |
Right-click the switch that is connected using the physical peer link and choose vPC Pairing from the drop-down list. The window to choose the peer appears.
You will get the following error when you choose a switch with the border gateway leaf role. <switch-name> has a Network/VRF
attached. Please detach the Network/VRF before vPC
Pairing/Unpairing |
||
Step 4 |
Check the Recommended column to see if pairing is possible. If the value is true, pairing is possible. You can pair switches even if the recommendation is false. However, you will get a warning or error during Save & Deploy. |
||
Step 5 |
Check the Use Virtual Peerlink check box. The Unpair icon changes to Save. |
||
Step 6 |
Click Save.
|
||
Step 7 |
In the Fabric Topology window, click Save & Deploy. The Config Deployment window appears. |
||
Step 8 |
Click the field against the switch in the Preview Config column. The Config Preview window appears for the switch. |
||
Step 9 |
View the vPC link details in the pending configuration and the side-by-side configuration. |
||
Step 10 |
Close the window. |
||
Step 11 |
Click the pending errors icon next to the Save & Deploy icon to view errors and warnings, if any. If you see any warnings that are related to TCAM, click the Resolve icon. A confirmation dialog box about reloading switches appears. Click OK. You can also reload the switches from Tabular view in the fabric topology window. The physical peer link between the peer switches turns red. Delete this link. The switches are connected only through a virtual peer link and are enclosed in a gray cloud. |
Converting a Virtual Peer Link to a Physical Peer Link
Before you begin
Procedure
Step 1 |
Choose Control > Fabrics. The Fabric Builder window appears. |
||
Step 2 |
Choose a fabric with the Easy_Fabric_11_1 or Easy_Fabric_eBGP fabric templates. |
||
Step 3 |
Right-click the switch that is connected through a virtual peer link and choose vPC Pairing from the drop-down list. The window to choose the peer appears.
|
||
Step 4 |
Uncheck the Use Virtual Peerlink check box. The Unpair icon changes to Save. |
||
Step 5 |
Click Save. |
||
Step 6 |
In the Fabric Topology window, click Save & Deploy. The Config Deployment window appears. |
||
Step 7 |
Click the field against the switch in the Preview Config column. The Config Preview window appears for the switch. |
||
Step 8 |
View the vPC peer link details in the pending configuration and the side-by-side configuration. |
||
Step 9 |
Close the window. |
||
Step 10 |
Click the pending errors icon next to the Save & Deploy icon to view errors and warnings, if any. If you see any warnings that are related to TCAM, click the Resolve icon. The confirmation dialog box about reloading switches appears. Click OK. You can also reload the switches from Tabular view in the fabric topology window. The virtual peer link, represented by a gray cloud, disappears and the peer switches are connected through a physical peer link. |
Advertising PIP on vPC
In the fabric settings, you can check the vPC advertise-pip check box to enable the Advertise PIP feature on all vPCs in a fabric. From Cisco DCNM Release 11.4(1), you can use the vpc_advertise_pip_jython policy to enable Advertise PIP feature on specific vPCs in a fabric.
Note the following guidelines:
-
If advertise-pip is not globally enabled or vPC peer is not using fabric peering, only then the vpc_advertise_pip_jython policy can be created on specific peers.
-
Enabling vpc advertise-pip doesn’t affect the current behavior.
-
Disabling advertise pip for a fabric doesn’t affect this policy.
-
Unpairing of switches deletes this policy.
-
You can manually delete this policy from the peer switch where it was created.
Procedure
Step 1 |
From the Fabric Builder window, click a fabric, and then right-click on a switch with vPC and select View/Edit Policies. |
||
Step 2 |
Click Add and select the vpc_advertise_pip_jython policy template and enter the mandatory parameters data.
|
||
Step 3 |
Click Save, and then deploy this policy. |
ThousandEyes Enterprise Agent
ThousandEyes Enterprise Agent collects network and application layer performance data when users access specific websites within monitored networks. It is used to run tests, check detailed aspects of network pathing and connectivity, status of network routing, monitor changes in intent, running configuration, and so on.
From Release 11.5(3), ThousandEyes Enterprise Agent is integrated with Cisco DCNM.
ThousandEyes Enterprise Agent is supported on Cisco Nexus 3000-R Series and Cisco Nexus 9000 Cloud Scale Series, with NX-OS version 9.3(7) and 10.2(1) and later releases.
This is supported with the following fabric templates:
-
Easy_Fabric_11_1
-
Easy_Fabric_eBGP
-
External_Fabric_11_1
-
LAN_Classic
You can configure global settings for ThousandEyes Enterprise Agent using Cisco DCNM
.The section includes the following:
Configuring TCAM and CoPP Policies
Ensure that you add relevant policies to Cisco Nexus 3000-R Series and Cisco Nexus 9000 Cloud Scale Series Switches before installing ThousandEyes Enterprise Agent feature on the switches.
To configure TCAM and CoPP policies on switches from the Cisco DCNM Web UI, perform the following steps:
Procedure
Step 1 |
From DCNM Web UI, choose Control > Fabric Builder, choose a fabric and click Tabular View in the Actions window. The Switches tab is displayed. |
||
Step 2 |
Select a single or multiple switches in the Switches tab and click the Policies button. |
||
Step 3 |
Click Add icon. |
||
Step 4 |
To add TCAM policies for Cisco Nexus 9000 EX, FX, and FX2 series switches perform following the steps:
|
||
Step 5 |
To add CoPP policies for Easy_Fabric_11_1 and Easy_Fabric_eBGP templates perform following the steps:
|
||
Step 6 |
To deploy the policy on all the supported switches and fabric templates, perform the following steps:
|
Performing ThousandEyes Enterprise Agent Actions
You can perform ThousandEyes Enterprise Agent action only for fabrics that are in managed mode.
Note |
Ensure that the TCAM and COPP policies are configured on switches, before installing ThousandEyes Enterprise Agent on it. |
To start, stop, install, or uninstall ThousandEyes Enterprise Agent using DCNM Web UI, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabric Builder. The Fabric Builder window appears. A rectangular box represents each fabric. |
||
Step 2 |
Choose a fabric and click Tabular View in the Actions window. The Switches tab is displayed. |
||
Step 3 |
Select single or multiple switches and click required action from ThousandEyes Agent drop- down list. You can perform following actions:
|
ThousandEyes Enterprise Agent Status
ThousandEyes Enterprise Agent status messages are as listed below:
-
NOT_INSTALLED - ThousandEyes Enterprise Agent is not installed on the switch.
-
RUNNING - ThousandEyes Enterprise Agent is active on the switch.
-
STOPPED - ThousandEyes Enterprise Agent has stopped on the switch.
-
UNSUPPORTED_VERSION - ThousandEyes Enterprise Agent is not supported with the switch NX-OS version.
-
UNSUPPORTED_PLATFORM - ThousandEyes Enterprise Agent is not supported on the selected switch platform.
-
NA - ThousandEyes Enterprise Agent global settings not configured on DCNM
-
Click ThousandEyes Status, to view information of ThousandEyes Enterprise Agent
The Detailed ThousandEyes Agent Information page appears.
-
Log Info tab displays the runtime agent status or error logs of the switch.
-
Sync Status tab displays deployed and expected settings details of switch.
-
DCNM indicates configuration mismatch (In-Sync, Out-Of-Sync) when ThousandEyes Enterprise Agent configuration is different from the effective configuration on DCNM at that instant. In case of configuration mismatch, you must uninstall, remove and install the ThousandEyes Enterprise Agent to make the configuration In-Sync.
Removing ThousandEyes Enterprise Agent
To remove the existing ThousandEyes Enterprise Agent entry in the ThousandEyes Enterprise portal, refer to instructions in Removing Old Agent Entries section.
To remove the existing ThousandEyes Enterprise Agent account group token from the switches on DCNM, perform the following steps:
Procedure
Step 1 |
From Cisco DCNM Web UI, choose Control > Fabric Builder. The Fabric Builder window appears. A rectangular box represents each fabric. |
Step 2 |
Choose a fabric and click Tabular View in the Actions window. The Switches tab is displayed. |
Step 3 |
Select the appropriate switches to remove ThousandEyes Enterprise Agent and click Play button (Execute Commands). The Execute Switch CLIs on Devices window appears. |
Step 4 |
Choose ThousandEyes_Agent_Identity_Delete from Template drop-down list and click Deploy. |
Viewing and Editing Policies
Cisco DCNM provides the ability to group a set of switches, and allows you to push a set of underlay configurations to the group. This release enables you to create a policy template, and apply it to multiple selected switches.
To view, add, deploy, or edit a policy, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabric Builder. |
||
Step 2 |
Select any available fabric, and then click Tabular view. |
||
Step 3 |
Select multiple switches in switches tab, and click View/Edit Policies.
|
Viewing Policies
Procedure
Step 1 |
Choose Control > Fabric Builder. |
||
Step 2 |
Select any available fabric, and then click Tabular view. |
||
Step 3 |
Select multiple switches in the switches tab and click View/Edit Policies. Policies are listed in view or edit policies table for multiple switches.
|
||
Step 4 |
Select a policy and click the View button to view its configs.
|
||
Step 5 |
In the View/Edit Policies window, click View All to view all the configurations pushed to the switches using policies. Go To: Select a device from this drop-down list to navigate to its starting config. This option is applicable only when you view policies for multiple devices. Include Policy ID: Select this check box to view policy IDs for all the policies. By default, this check box is selected. |
Adding a Policy
Procedure
Step 1 |
Choose Control > Fabric Builder. |
Step 2 |
Select any available fabric, and then click Tabular view. |
Step 3 |
Select a single or multiple switches in the Switches tab, and click the View/Edit Policies button. |
Step 4 |
Click the Add icon. |
Step 5 |
Select a policy template and enter the mandatory parameters data and click Save. PTI is added per each device based on n-number of devices selection. Policy: Select a policy from this drop-down list. Priority: Specify a priority for the policy. The applicable values are from 1 to 1000. The default value is 500. The lower number in the Priority field means that there is a higher priority for the generated configuration and POAP startup-configuration. For example, features are 50, route-maps are 100, and vpc-domain is 200. Description: (Optional) Specify a description for the policy. This field is used to differentiate multiple freeform policies. The Description column is added in the View/Edit Policies window, which you can use to filter or find policies based on description. |
Deploying Policies
Procedure
Step 1 |
Choose Control > Fabric Builder. |
Step 2 |
Select any available fabric, and then click Tabular view. |
Step 3 |
Select multiple switches in the switches tab, and click the View/Edit Policies button. |
Step 4 |
Select multiple polices, and then click Push Config. The selected PTI’s configs are pushed to the group of switches.
|
Editing a Policy
Note |
Multiple policy editing is not supported. |
Procedure
Step 1 |
Choose Control > Fabric Builder. |
||
Step 2 |
Select any available fabric, and then click Tabular view. |
||
Step 3 |
Select multiple switches in the switches tab, and click the View/Edit Policies button.
|
||
Step 4 |
Select a PTI, click Edit to modify the required data, and then click Save to save the PTI. |
||
Step 5 |
Select a PTI, click Edit to modify the required data, and then click Push Config to push the policy config to the device.
|
Current Switch Configuration
Procedure
Step 1 |
Choose Control > Fabric Builder. |
||
Step 2 |
Select any available fabric, and then click Tabular view. |
||
Step 3 |
Select multiple switches in the switches tab, and click View/Edit Policies. |
||
Step 4 |
Click Current Switch Config. The current switch configuration appears in the Running Config dialog box.
|
Retrieving the Authentication Key
Retrieving the 3DES Encrypted OSPF Authentication Key
-
SSH into the switch.
-
On an unused switch interface, enable the following:
config terminal feature ospf interface Ethernet1/1 no switchport ip ospf message-digest-key 127 md5 ospfAuth
In the example, ospfAuth is the unencrypted password.
Note
This Step 2 is needed when you want to configure a new key.
-
Enter the show run interface Ethernet1/1 command to retrieve the password.
Switch # show run interface Ethernet1/1 interface Ethernet1/1 no switchport ip ospf message-digest key 127 md5 3 sd8478f4fsw4f4w34sd8478fsdfw no shutdown
The sequence of characters after md5 3 is the encrypted password.
-
Update the encrypted password into the OSPF Authentication Key field.
Retrieving the Encrypted IS-IS Authentication Key
To get the key, you must have access to the switch.
-
SSH into the switch.
-
Create a temporary keychain.
config terminal key chain isis key 127 key-string isisAuth
In the example, isisAuth is the plaintext password. This will get converted to a Cisco type 7 password after the CLI is accepted.
-
Enter the show run | section “key chain” command to retrieve the password.
key chain isis key 127 key-string 7 071b245f5a
The sequence of characters after key-string 7 is the encrypted password. Save it.
-
Update the encrypted password into the ISIS Authentication Key field.
-
Remove any unwanted configuration made in Step 2.
Retrieving the 3DES Encrypted BGP Authentication Key
-
SSH into the switch and enable BGP configuration for a non-existent neighbor.
Note
Non-existent neighbor configuration is a temporary BGP neighbor configuration for retrieving the password.
router bgp neighbor 10.2.0.2 remote-as 65000 password bgpAuth
In the example, bgpAuth is the unencrypted password.
-
Enter the show run bgp command to retrieve the password. A sample output:
neighbor 10.2.0.2 remote-as 65000 password 3 sd8478fswerdfw3434fsw4f4w34sdsd8478fswerdfw3434fsw4f4w3
The sequence of characters after password 3 is the encrypted password.
-
Update the encrypted password into the BGP Authentication Key field.
-
Remove the BGP neighbor configuration.
Retrieving the Encrypted BFD Authentication Key
-
SSH into the switch.
-
On an unused switch interface, enable the following:
switch# config terminal switch(config)# int e1/1 switch(config-if)# bfd authentication keyed-SHA1 key-id 100 key cisco123
In the example, cisco123 is the unencrypted password and the key ID is 100.
Note
This Step 2 is needed when you want to configure a new key.
-
Enter the show running-config interface command to retrieve the key.
switch# show running-config interface Ethernet1/1 interface Ethernet1/1 description connected-to- switch-Ethernet1/1 no switchport mtu 9216 bfd authentication Keyed-SHA1 key-id 100 hex-key 636973636F313233 no ip redirects ip address 10.4.0.6/30 no ipv6 redirects ip ospf network point-to-point ip router ospf 100 area 0.0.0.0 no shutdown
The BFD key ID is 100 and the encrypted key is 636973636F313233.
-
Update the key ID and key in the BFD Authentication Key ID and BFD Authentication Key fields.
Custom Maintenance Mode Profile Policy
When you place a switch in maintenance mode using DCNM, only a fixed set of BGP and OSPF isolate CLIs are configured in the maintenance mode profile. Starting from Cisco DCNM Release 11.3(1), you can create a custom_maintenance_mode_profile PTI with customized configurations for maintenance mode and normal mode profile, deploy the PTI to the switch, and then move the switch to maintenance mode.
Creating and Deploying a Custom Maintenance Mode Profile Policy
Procedure
Step 1 |
Select Control>Fabric Builder, click Tabular View, and select a switch in the Name column or select Control>Fabric Builder and right-click the switch. |
Step 2 |
Click View/Edit Policies and click on + to add a new policy. The Add Policy window comes up. |
Step 3 |
Select custom_maintenance_mode_profile from the Policy dropdown list. |
Step 4 |
Fill in the Maintenance mode profile contents with the desired configuration CLIs. Example:
Fill in the Normal mode profile contents with the desired configuration CLIs. Example:
|
Step 5 |
Click Save. |
Step 6 |
Right-click the switch in the Fabric Builder window and select Deploy Config. Verify the configuration in the Pending Config window and then deploy the configuration to the switch. |
Step 7 |
Then, right-click the switch and select Modes>Maintenance Mode to move the switch to maintenance mode. |
Deleting a Custom Maintenance Mode Profile Policy
Procedure
Step 1 |
The switch has to be moved to active/operational or normal mode before deleting the custom maintenance mode profile policy. To do this, right-click the switch in the Fabric Builder window and select Modes>Active/ Operational Mode. |
Step 2 |
After the switch has been moved to active/operational or normal mode, click Tabular View in the Fabric Builder window, and select the switch in the Name column or right-click the switch in the Fabric Builder window. |
Step 3 |
Click View/Edit Policies, and select the custom_maintenance_mode_profile policy that has to be deleted. |
Step 4 |
Click X to delete the policy. |
Step 5 |
Right-click the switch in the Fabric Builder window and select Deploy Config. Verify the configuration in the Pending Config window and deploy the configuration to the switch.
|
Return Material Authorization (RMA)
This section describes how to replace a physical switch in a Fabric when using Cisco DCNM Easy Fabric mode.
Prerequisites
-
Ensure that the fabric is up and running with minimal disruption while replacing the switch.
-
To use the POAP RMA flow, configure the fabric for bootstrap (POAP).
-
Perform save and deploy more than once, if needed, to copy the FEX configurations for the RMA of switches that have FEX deployed.
Guidelines and Limitations
-
To replace the switch, remove the old switch from the fabric and discover the new switch in the fabric. For example, if you want to replace a Cisco Nexus 9300-EX switch with a Cisco Nexus 9300-FX switch, remove the 9300-EX switch from the fabric followed by discovering the 9300-FX switch in the same fabric.
-
When GIR is enabled before upgrading Cisco Nexus 7000 Series switches, DCNM pushes the system mode maintenance command to the switches when the DCNM RMA procedure is initiated. This command applies the configuration that is present in the default maintenance mode profile to the switches. For more information on performing Graceful Insertion and Removal (GIR) on the Cisco Nexus 7000 Series switches, refer Configuring GIR.
If a switch is in maintenance mode, to initiate copy run and start configuration, navigate to
and reload the switch.
POAP RMA Flow
Procedure
Step 1 |
Choose Control > Fabric Builder. |
Step 2 |
Click the Fabric where you want to perform RMA. |
Step 3 |
Move the device into maintenance mode. To move a device into maintenance mode, right-click on the device, and then choose Modes > Maintenance Mode. |
Step 4 |
Physically replace the device in the network. Physical connections should be made in the same place on the replacement switch as they existed on the original switch. |
Step 5 |
Provision RMA flow and select the replacement device. |
Step 6 |
The Provision RMA UI will show the replacement device 5-10 minutes after it is powered on. |
Step 7 |
Select the correct replacement device and click Swap Switch. This begins POAP with the full “expected” configuration for that device. Total POAP time is generally around 10-15 minutes. |
Manual RMA Flow
Use this flow when “Bootstrap” is not possible (or not desired), including cases that are IPv6 only for the initial Cisco DCNM 11.0(1) release.
Procedure
Step 1 |
Place the device in maintenance mode (optional). |
Step 2 |
Physically replace the device in the network. |
Step 3 |
Log in through Console and set the Management IP and credentials. |
Step 4 |
The Cisco DCNM rediscovers the new device (or you can manually choose Discovery > Rediscover). |
Step 5 |
Deploy the expected configuration using Deploy. |
Step 6 |
Depending on the configuration, if breakout ports or FEX ports are in use, you have to deploy again to completely restore the configuration. |
Step 7 |
After a successful deployment, and the device is “In-Sync,” you must move the device back to Normal Mode. |
RMA for User with Local Authentication
Note |
This task is only applicable to non-POAP switches. |
Use the following steps to perform RMA for a user with local authentication:
Procedure
Step 1 |
After the new switch comes online, SSH into the switch and reset the local user passwords with the cleartext password using the “username” command. Reset the local user passwords to resync the SNMP password. The password is stored in the configuration file in a nontransferable form. |
Step 2 |
Wait for the RMA to complete. |
Step 3 |
Update Cisco DCNM switch_snmp_user policy for the switch with the new SNMP MD5 key from the switch. |
Interfaces
The Interfaces option displays all the interfaces that are discovered for the switch, Virtual Port Channels (vPCs), and intended interfaces missing on the device.
You can use the following functions:
-
Create, deploy, view, edit and delete a port channel, vPC, Straight-through FEX, Active-Active FEX, loopback, and subinterface.
Note
-
The following features are unsupported for the brownfield migration of switches with the Cisco NX-OS Release 7.0(3)I4(8b) and 7.0(4)I4(x) images:
-
FEX on switches other than Cisco Nexus 9300 Series switches and Cisco Nexus 9500 Series switches with X9500 line cards
-
AA-FEX
For information about the platform support for FEX, refer to your platform and NX-OS documentation to check the feature compatibility.
-
-
To edit interfaces associated with fabric links such as intra-fabric links and inter-fabric links, see Editing Interfaces Associated with Links.
-
The flowcontrol or priority-flow-control config is not supported for HIF ports or PO with HIF ports as members.
-
-
Create tunnel interfaces for Cisco Cloud Services Router 1000v Series (Cisco CSR 1000v Series).
-
Create breakout and unbreakout ports.
-
Shut down and bring up interfaces.
-
Rediscover ports and view interface configuration history.
-
Apply host policies on interfaces and vPCs. For example, int_trunk_host_11_1, int_access_host_11_1, and so on.
-
View interface information such as its admin status, operation status, reason, policy, speed, MTU, mode, VLANs, IP/Prefix, VRF, port channel, and the neighbor of the interface.
Note
-
The Neighbor column provides details of connected switches that are discovered, intent links, and Virtual Machine Manager (VMM) connectivity. You can navigate to the Switch dashboard of the corresponding switch by clicking it. However, intent links and VMM links aren’t hyperlinked and you cannot navigate to the corresponding Switch dashboard.
-
Click the graph icon in the Name column to view the interface performance chart for the last 24 hours. However, note that performance data for VLAN interfaces that are associated with overlay networks is not displayed in this chart.
The Status column displays the following statuses of an interface:
-
Blue: Pending
-
Green: In Sync/Success
-
Red: Out-of-Sync/Failed
-
Yellow: In Progress
-
Grey: Unknown/NA
-
-
If an interface is created out-of-band, you need to perform fabric resync or wait for Config Compliance poling before this interface can be deleted. Otherwise, Config Compliance does not generate the correct diff.
However, you cannot add or edit interfaces for ASR 9000 Series Routers and Arista switches.
You can filter and view information for any of the given fields (such as Device Name). The following table describes the buttons that appear on this page.
Note |
|
Field |
Description |
---|---|
Add |
Allows you to add a logical interface such as a port channel, vPC, Straight-through FEX, Active-Active FEX, loopback and subinterface. |
Breakout, Unbreakout |
Allows you to breakout an interface or unbreakout interfaces that are in breakout state. |
Edit |
Allows you to edit and change policies that are associated with an interface. |
Delete |
Allows you to delete a logical interface that is created from the Interfaces screen. An interface having a policy that is attached from an overlay and underlay cannot be deleted. |
No Shutdown |
Allows you to enable an interface (no shutdown or admin up). |
Shutdown |
Allows you to shut down the interface. |
Show |
Allows you to display the interface show commands. A show command requires show templates in the template library. |
Rediscover |
Allows you to rediscover or recalculate the compliance status on the selected interfaces. |
Interface History |
Allows you to display the interface deployment history details. |
Deploy |
Allows you to deploy or redeploy saved interface configurations. |
The following table describes the different user roles and the operations these roles support in the Interfaces window from Cisco DCNM Release 11.4(1).
Operations | User Roles | ||
---|---|---|---|
network-admin | network-operator | network-stager | |
Add | Save, Preview, Deploy | Blocked | Save, Preview |
Breakout | Supported | Blocked | Blocked |
Unbreakout | Supported | Blocked | Blocked |
Edit | Save, Preview, Deploy | Preview | Save, Preview |
Delete | Save, Preview, Deploy | Blocked | Save, Preview |
Shutdown | Save, Preview, Deploy | Blocked | Save, Preview |
No Shutdown | Save, Preview, Deploy | Blocked | Save, Preview |
Show | Supported | Supported | Supported |
Rediscover | Supported | Supported | Supported |
Deploy | Preview, Deploy | Blocked | Blocked |
The following table describes the new user role access-admin operations support in the host facing port of Interfaces window from Cisco DCNM Release 11.5(1).
Operations | User Roles | ||
---|---|---|---|
access-admin | |||
Add | Save, Preview, Deploy | ||
Breakout | Blocked | ||
Unbreakout | Blocked | ||
Edit | Save, Preview, Deploy
|
||
Delete | Save, Preview, Deploy | ||
Shutdown | Save, Preview, Deploy | ||
No Shutdown | Save, Preview, Deploy | ||
Show | Supported | ||
Rediscover | Supported | ||
Deploy | Preview, Deploy |
From Cisco DCNM Release 11.4(1), you can disable deployments, or freeze, a fabric in DCNM as a network administrator. However, you cannot perform all actions when you freeze the fabric or if the fabric is in monitor mode.
The following table describes the actions you can perform when you freeze a fabric and when you enable the monitor mode for a fabric.
Operations | DCNM Mode | |
---|---|---|
Freeze Mode | Monitor Mode | |
Add | Save, Preview | Blocked |
Breakout | Blocked | Blocked |
Unbreakout | Blocked | Blocked |
Edit | Save, Preview | Blocked |
Delete | Save, Preview | Blocked |
Shutdown | Save, Preview | Blocked |
No Shutdown | Save, Preview | Blocked |
Show | Supported | Supported |
Rediscover | Supported | Supported |
Deploy | Blocked | Blocked |
The buttons for the associated operations are grayed out accordingly.
If you perform admin operations (shutdown/no shutdown) on SVI, which is part of a config profile, successive Save & Deploy operations generate no interface vlan command.
For SVI with no policy, on performing admin operation, that is, shutdown/no shutdown command pushed from Interface Manager, int_vlan_admin_state policy is associated with the SVI.
For example, create and deploy the SVI from switch_freeform.
interface vlan1234
description test
no shutdown
no ip redirects
no ipv6 redirects
If you shutdown the SVI from interface manager, the int_vlan_admin_state policy is associated with the SVI.
Pending diff is shown as:
interface Vlan1234
shutdown
no ip redirects
no ipv6 redirects
description test
no shutdown
Remove the no shutdown CLI from the free-form config.
If the user has performed admin operation on SVI, device will have interface in running config. Therefore, post network detach interface vlan will be still present and interface will be discovered. You need to manually delete the interface from Interface Manager.
This section contains the following:
Adding Interfaces
Procedure
Step 1 |
Choose Control > Interfaces. You see the Scope option at the top right. If you want to view interfaces for a specific fabric, select the fabric window from the list. |
||
Step 2 |
Click Add to add a logical interface. The Add Interface window appears. |
||
Step 3 |
In the Type drop-down list, choose the type of the interface.
|
||
Step 4 |
In the Select a Device field, choose the device. Devices are listed based on the fabric and interface type. External fabric devices aren’t listed for ST FEX and AA FEX. In the case of vPC or Active to Active FEX, select the vPC switch pair. |
||
Step 5 |
Enter the ID value in the respective interface ID field (Port-channel ID, vPC ID, Loopback ID and Subinterface ID) that is displayed, based on the selected interface. You can override this value. The new value is used only if it’s available in the Resource Manager pool. Else, it results in an error. |
||
Step 6 |
In the Policy field, select the policy to apply on an interface. The field only lists the Interface Python Policy with tag interface_edit_policy and filtered based on the interface type. You must not create a _upg interface policy. For example, you shouldn’t create a policy using the vpc_trunk_host_upg, port_channel_aa_fex_upg, port_channel_trunk_host_upg, and trunk_host_upg options.
|
||
Step 7 |
Enter values in the required fields under the General tab. The fields vary according to the interface type you choose.
|
||
Step 8 |
Click Save to save the configurations.
Only saved configurations are pushed to the device. While adding the interface, you can only modify the policy attribute after the first save. If you try to use an ID that is already used, you encounter the Resource could not be allocated error. |
||
Step 9 |
(Optional) Click the Preview option to preview the configurations to be deployed. |
||
Step 10 |
Click Deploy to deploy the specified logical interface. The newly added interface appears in the screen. |
Breakout
Click the drop-down arrow next to the Breakout icon to display a list of the available breakout options. The available options are 10g-4x, 25g-4x, 50g-2x, 50g-4x, 100g-2x, 100g-4x, 200g-2x, and Unbreakout. Choose the required option.
Editing Interfaces
To edit the interfaces from the Cisco DCNM Web UI, perform the following steps:
Note |
The Edit Interface allows you to change the policy and add or remove an interface from a port channel or vPC. |
Procedure
Step 1 |
Choose Control > Interfaces. You can break out and unbreak out an interface by using the breakout option at the top left part of the screen. |
Step 2 |
Select the interface check box to edit an interface or vPC. Select corresponding check boxes for editing multiple interfaces. You cannot edit multiple port channels and vPC. You cannot edit interfaces of different types at the same time. |
Step 3 |
Click Edit to edit an interface. The variables that are shown in the Edit Configuration window are based on the template and its policy. Select the appropriate policy. Preview the policy, save it and deploy the same. This window lists only Interface Python Policy with the tag interface_edit_policy and filtered based on the interface type. In a vPC setup, the two switches are in the order the switch names are displayed in the edit window. For example, if Switch Name is displayed as LEAF1:LEAF2, then Leaf1 is peer switch one and Leaf2 is peer switch two. During overlay network deployment on switches, the network can be associated with trunk interfaces. The trunk interface to network association is reflected in the Interfaces screen. You can update such interfaces. For interface policies that are not created from the Control > Interfaces screen, you can edit some configurations but not change the policy itself. The policy and fields that cannot be edited are grayed out. The following are some examples of policies that cannot be edited:
|
Editing Interfaces Associated with Links
There are two types of links, namely intra-fabric links and inter-fabric links. As the name implies, intra-fabric links are set up between devices within the same Easy fabric and are typically used for spine-leaf connectivity. Inter-fabric links are set up between the Easy fabric, and typically other external or Easy fabrics. They are used for external WAN and/or DCI connectivity. A policy is associated with each link that effectively states the configuration that is applied to both ends of the link. In other words, the link policy becomes the parent of the individual child interface policies that are associated with the two interfaces that form the link. In this scenario, you must edit the link policy to edit the interface policy fields such as description, IP address, and any per interface freeform config. The following procedure shows how to edit the interfaces associated with links:
Procedure
Step 1 |
Choose Control > Fabric Builder, and select the fabric containing the link. |
Step 2 |
Click Tabular view in the Actions panel. A window with the Switches and Links tabs appears. |
Step 3 |
Click the Links tab. |
Step 4 |
Select the link that you want to edit and click the Update Link icon. Update the link based on your requirements and click Save. |
Deleting Interfaces
To delete the interfaces from the Cisco DCNM Web UI, perform the following steps:
Note |
This option allows you to delete only logical ports, port channels, and vPCs. You can delete the interface if it does not have overlay or underlay policy attached. When a port channel or vPC is removed, the corresponding member ports get the default policy associated. The Default Policy can be configured in server.properties file. |
Procedure
Step 1 |
Choose Control > Interfaces. |
Step 2 |
Select the interfaces. |
Step 3 |
Click Delete. You cannot delete logical interfaces created in the fabric underlay. |
Step 4 |
Click Save. |
Step 5 |
(Optional) Click Preview to view all the changes before deleting the interface. The deletion will be highlighted in red colour with strikethrough under the Expected Config tab. |
Step 6 |
Click Deploy to delete the interface. |
Shutting Down and Bringing Up Interfaces
Procedure
Step 1 |
Choose Control > Interfaces. |
Step 2 |
Select the interfaces that you want to shut down or bring up. |
Step 3 |
Click Shutdown to disable the selected interfaces. For example, you may want to isolate a host from the network or a host that is not active in the network. A confirmation window appears where you can save, preview, and deploy the changes. Click Save to preview of deploy the changes. |
Step 4 |
Click No Shutdown to bring up the selected interfaces. A confirmation window appears where you can save, preview, and deploy the changes. Click Save to preview or deploy the changes. |
Viewing Interface Configuration
Procedure
Step 1 |
Choose Control > Interfaces. Select the interface whose configurations you want to view. |
Step 2 |
In the Interface Show Commands window, select the action from the Show drop-down box and click Execute. The interface configurations are displayed in the Output section, at the right of the screen. For Show commands, you must have corresponding show templates for interface or interface sub types like port channel or vPC, defined in the Template Library. |
Rediscovering Interfaces
Procedure
Step 1 |
Choose Control > Interfaces. |
Step 2 |
Select the interfaces that you want to rediscover. |
Step 3 |
Click Rediscover to rediscover the selected interfaces. For example, after you edit or enable an interface, you can rediscover the interface. |
Viewing Interface History
Procedure
Step 1 |
Choose Control > Interfaces. |
Step 2 |
Select the interface. |
Step 3 |
Click Interface History to view the configuration history on the interface. |
Step 4 |
Click Status to view each command that is configured for that configuration instance. |
Deploying Interface Configurations
Procedure
Step 1 |
Choose Control > Interfaces. |
||
Step 2 |
Choose an interface you want to deploy.
|
||
Step 3 |
Click Deploy to deploy or redeploy configurations that are saved for an interface. After you deploy the interface configuration, the interface status information is updated. However, the overall switch-level state may be in the pending state, which is in blue. The overall switch-level state goes to the pending state whenever there is a change in intent from any module, such as interface, link, policy template update, top-down, or so on. In the pending state, a switch may have pending configurations or switch-level recomputation. The switch-level recomputation occurs when:
Preview or deploy the switches to review their state and to understand the root cause of their pending state. Save and deploy for a fabric-wide recomputation. Click Preview to preview the configurations before you click Deploy. |
Creating External Fabric Interfaces
You can add and edit port channel, vPC, subinterface, and loopback interfaces for external fabric devices. You cannot add Straight-through FEX and Active-Active FEX functions.
The Breakout port function is only supported for the Cisco Nexus 9000, 3000, and 7000 Series Switches in the external fabric.
When you add an interface to an external fabric device, the Resource Manager is not in sync with the device. So, ensure that the value populated in the ID field (Port-channel ID, vPC ID, Loopback ID, etc) is not previously configured on the switch.
If you want to configure a portchannel in the external fabric, you should add and deploy the feature_lacp policy on the switches where the portchannel will be configured.
When an external fabric is set to Fabric Monitor Mode Only, you cannot deploy configurations on its switches. If you click Save & Deploy in the fabric topology screen, it displays an error message. However, the following settings (available when you right-click the switch icon) are allowed:
vPC pairing - You can designate a vPC switch pair, but it is only for reference.
View/edit policy - You can add a policy but you cannot deploy it on the switch.
Manage interfaces – You can only create intent for adding interfaces. If you try to deploy, edit, or delete interfaces, it results in an error message.
Interface Groups
From Cisco DCNM Release 11.5(1), you can create an interface group that allows grouping of host-facing interfaces at a fabric level. Specifically, you can create an interface group for physical Ethernet interfaces, L2 port-channels, and vPCs. You can attach or unattach multiple overlay networks to the interfaces in an interface group.
Guidelines
-
Interface groups are only supported for the fabrics with the Easy_Fabric_11_1 template.
-
An interface group is specific to a fabric. For example, consider two fabrics: Fab1 and Fab 2. The interface group IG1 in Fab1 isn’t applicable to Fab 2.
-
An interface group can only have interfaces of a certain type. For example, you need three separate interface groups if you want to group three types of interfaces such as IG1 for physical Ethernet trunk interfaces, IG2 for L2 trunk port-channels, and IG3 for vPC host trunk ports.
-
An interface group can also be created using preprovisioned interfaces.
-
Interface groups are limited to switches with the leaf role. They aren’t supported for other roles such as Border, BGW, and other related variants.
-
For L2 port-channels and vPCs that are part of an interface group, they can’t be deleted until they are de-associated from the interface group even if there are no networks associated with the interface group. Similarly, a trunk port that has no overlay networks but is part of an IG can’t be converted to an access port. In other words, you can’t change policies for interfaces that are part of an interface group. However, you can edit certain fields for policies.
-
For L4-L7 services configuration on leaf switches, trunk ports that are used for services attachment can’t be part of interface groups.
-
When you perform a per fabric backup of an easy fabric, if there are interface groups created in that fabric, all the associated interface group state is backed up.
-
If an easy fabric contains an interface group, then this fabric can’t be imported into the MSO. Similarly, if an easy fabric has been added to the MSO, you can’t create interface groups for interfaces that belong to switches in the easy fabric.
-
The Interface Group button is enabled only for Admin and Stager users. For all other users, this button is disabled.
-
The Interface Group button is disabled in the following circumstances:
-
Select Data center from the SCOPE drop-down list.
-
Select a fabric without any switches.
-
Select any other interface apart from vPC, Port-channel, and Ethernet.
-
If the interface has a policy attached from another source, for example:
-
If the interface is member of a port-channel or vPC.
-
If the port-channel is member of vPC.
-
If the interface has a policy from underlay or links.
-
Note
If you select different types of interfaces, the Interface Group button is enabled. However, when you try to create or save different types of interfaces to an interface group, an error is displayed.
-
Creating an Interface Group
Procedure
Step 1 |
From DCNM, navigate to Control > Fabrics > Interfaces. |
||
Step 2 |
From the SCOPE drop-down list, select a fabric. |
||
Step 3 |
Select the interfaces that have to be grouped and click Interface Group. |
||
Step 4 |
4. In the Edit Interface Group window, create a custom interface group by entering an interface group name in the Select Interface Group field and click Create custom. An interface group name can have a maximum length of 64 characters. If you have already created an interface group, select it from the Select Interface Group drop-down list. Also, if an interface is already part of an interface group, you can move it to a different interface group by selecting the new group from the Select Interface Group drop-down list.
You can create interface groups from either the Interfaces window or the Networks window. For more information, see Attaching Networks to an Interface Group. |
||
Step 5 |
Click Save. In the Interfaces window, you can see the interface group name under the Interface Group column. |
Removing Interfaces from an Interface Group
Procedure
Step 1 |
From DCNM, navigate to Control > Fabrics > Interfaces. |
Step 2 |
From the SCOPE drop-down list, select a fabric. |
Step 3 |
Select the interfaces to disassociate from an interface group and click Interface Group. |
Step 4 |
In the Edit Interface Group window, make sure that nothing is selected in the Select Interface Group drop-down list, and click Clear. A dialog box pops up asking whether you want to clear all the associated interfaces. Click Yes to proceed. Note that if there are any networks attached to these interfaces, they are detached as well when you click Clear. |
Attaching Networks to an Interface Group
Procedure
Step 1 |
From DCNM, navigate to Control > Fabrics > Networks. |
||
Step 2 |
From the SCOPE drop-down list, select a fabric. |
||
Step 3 |
In the Networks window, select the networks that you need to attach to an interface group and click Interface Group.
|
||
Step 4 |
In the Interface Groups window, you can perform the following:
|
||
Step 5 |
Click Continue and click Save & Deploy to deploy the selected networks on the switches. |
Unattaching a Network from an Interface Group
This procedure shows how to unattach a network from an interface group in the Networks window. Also, you can unattach networks when you remove an interface from an interface group in the Interfaces window. For more information, see Removing Interfaces from an Interface Group.
Procedure
Step 1 |
1. From DCNM, navigate to Control > Fabrics > Networks. |
Step 2 |
From the SCOPE drop-down list, select a fabric. |
Step 3 |
In the Networks window, select the networks that you need to unattach to an interface group and click Interface Group. |
Step 4 |
In the Interface Groups window, select the interface group from the Select Interface Group drop-down list and click Clear to unattach a network. |
Step 5 |
(Optional) Navigate to Control > Fabrics > Interfaces. Under the Overlay Network column, you can see the unattached network in the red color for the corresponding interface. Click the network to view the expected config that is struck through. |
Step 6 |
Navigate to the Fabric Builder or Networks window, and click Save & Deploy. |
Deleting an Interface Group
An interface group is automatically deleted when it’s not in use. DCNM performs an implicit delete of an interface group if there are no interfaces and no networks mapped to the interface group. This check is performed whenever you click the Clear button in the Edit Interface Group window. There may be exception scenarios where you need to clean up the interface groups explicitly.
For example, you create an interface group storageIG and add an interface to it. Later, you want to change the interface mapping to another group. Therefore, you select the interface and click Interface Group to open the Edit Interface Group window. Select the other interface group named diskIG. Now, the storageIG interface group doesn’t have any associated member interfaces or networks. In this case, perform the following steps:
Procedure
Step 1 |
Select an interface that doesn’t belong to an interface group. |
Step 2 |
Click Interface Group to open the Edit Interface Group window. |
Step 3 |
Select the StorageIG interface group from the Select Interface Group drop-down list. |
Step 4 |
Click Clear. |
Creating and Deploying Networks and VRFs
The steps for overlay networks and VRFs provisioning are:
-
Create networks and VRFs for the fabric.
-
Deploy the networks and VRFs on the fabric switches.
Note |
The undeployment and deletion of overlay networks and VRFs are explained after the explanation of deployment. Finally, creation of external fabrics and fabric extensions from VXLAN to external fabrics are documented. |
For information about creating interface groups and attaching networks, see Interface Groups.
You can navigate to the networks and VRFs window through any of the following options:
-
From the home page: Click the Networks & VRFs button in the Cisco DCNM Web UI landing page.
-
From the Control menu: From the home page of the Cisco DCNM Web UI, choose Control > Fabrics > Networks to navigate to the Networks window. Choose Control > Fabrics > VRFs to navigate to the VRFs window.
-
From a fabric topology window: Right-click anywhere in the fabric topology window. Choose Overlay View > VRF View or Overlay View > Network View, accordingly. This option is applicable only for switch fabrics, easy fabrics, and MSD fabrics.
You can toggle between the network view and VRF view in both the windows by clicking the VRF View or Network View button. When you are in the networks or VRFs window, ensure you choose the appropriate fabric from the Scope drop-down list before you create any networks or VRFs.
Viewing Networks and VRFs for a Fabric
-
Click Control > Networks from the main menu.
The Networks screen comes up. The SCOPE drop down box (at the top right part of the screen) lists all fabrics managed by the DCNM instance, in alphabetical order. You can choose the correct fabric from SCOPE. When you select a fabric, the Networks screen refreshes and lists networks of the selected fabric.
-
Click Control > VRFs from the main menu.
The VRFs screen comes up. The SCOPE drop down box (at the top right part of the screen) lists all fabrics managed by the DCNM instance, in alphabetical order. You can choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists VRFs of the selected fabric.
Note |
The Networks or VRFs windows are applicable only for the Easy or MSD fabrics. |
Creating Networks for the Standalone Fabric
-
Click Control > Networks (under Fabrics submenu).
The Networks screen comes up.
-
Choose the correct fabric from SCOPE. When you select a fabric, the Networks screen refreshes and lists networks of the selected fabric.
-
Click the + button at the top left part of the screen (under Networks) to add networks to the fabric. The Create Network screen comes up. Most of the fields are autopopulated.
The fields in this screen are:
Network ID and Network Name: Specifies the Layer 2 VNI and name of the network. The network name should not contain any white spaces or special characters except underscore (_) and hyphen (-). The corresponding Layer 3 VNI (or VRF VNI) is generated along with VRF creation.
VRF Name: Allows you to select the Virtual Routing and Forwarding (VRF).
When no VRF is created, this field appears blank. If you want to create a new VRF, click the + button. The VRF name should not contain any white spaces or special characters except underscore (_), hyphen (-), and colon (:).
Layer 2 Only: Specifies whether the network is Layer 2 only.
Network Template: A universal template is autopopulated. This is only applicable for leaf switches.
Network Extension Template: A universal extension template is autopopulated. This allows you to extend this network to another fabric. The methods are VRF Lite, Multi Site, and so on. The template is applicable for border leaf switches and BGWs.
VLAN ID: Specifies the corresponding tenant VLAN ID for the network.
The VLAN ID default range is 2 to 3967. From DCNM Release 11.5(2), you can use a VLAN range greater than default value 3967. The reserved VLAN range must be set to a different range. In switch command enter “system vlan <vlan> reserve”. Save the configuration to startup configuration and reload the switch for the new reserved VLAN range to reflect.
From Cisco DCNM Web UI, choose Administration > DCNM Server > Server Properties, enter the value as 4094 for RM.TOP_DOWN_NETWORK_VLAN.MAX and RM.TOP_DOWN_VRF_VLAN.MAX, click Apply Changes and then restart DCNM. Once the DCNM is up, you can create VRF and network using the VLAN value greater than 3967.
Network Profile section contains the General and Advanced tabs.
General tab
IPv4 Gateway/NetMask: Specifies the IPv4 address with subnet.
Note
If the same IP address is configured in the IPv4 Gateway and IPv4 Secondary GW1 or GW2 fields of the network template, DCNM does not show an error, and you will be able to save this configuration.
However, after the network configuration is pushed to the switch, it would result in a failure as the configuration is not allowed by the switch.
IPv6 Gateway/Prefix: Specifies the IPv6 address with subnet.
Specify the anycast gateway IP address for transporting the L3 traffic from a server belonging to MyNetwork_30000 and a server from another virtual network. By default the anycast gateway IP address is the same for MyNetwork_30000 on all switches of the fabric that have the presence of the network.
VLAN Name - Enter the VLAN name.
Interface Description: Specifies the description for the interface. This interface is a switch virtual interface (SVI).
MTU for the L3 interface - Enter the MTU for Layer 3 interfaces.
IPv4 Secondary GW1 - Enter the gateway IP address for the additional subnet.
IPv4 Secondary GW2 - Enter the gateway IP address for the additional subnet.
Advanced tab: Optionally, specify the advanced profile settings by clicking the Advanced tab:
ARP Suppression – Select the checkbox to enable the ARP Suppression function.
Ingress Replication - The checkbox is selected if the replication mode is Ingress replication.
Note
Ingress Replication is a read-only option in the Advanced tab. Changing the fabric setting updates the field.
Multicast Group Address- The multicast IP address for the network is autopopulated.
Multicast group address is a per fabric instance variable. The number of underlay multicast groups supported is only 128. If all networks are deployed on all switches, you need not use a different multicast group per L2 VNI or a network. Therefore, multicast group for all networks in a fabric remains same. If a new multicast group address is required, you can generate it by clicking the Generate Multicast IP button.
DHCPv4 Server 1 - Enter the DHCP relay IP address of the first DHCP server.
DHCPv4 Server 2 - Enter the DHCP relay IP address of the next DHCP server.
DHCPv4 Server VRF- Enter the DHCP server VRF ID.
Loopback ID for DHCP Relay interface (Min:0, Max:1023) - Specifies the loopback ID for DHCP relay interface.
Routing Tag – The routing tag is autopopulated. This tag is associated with each gateway IP address prefix.
TRM enable – Select the check box to enable TRM.
For more information, see Overview of Tenant Routed Multicast.
L2 VNI Route-Target Both Enable - Select the check box to enable automatic importing and exporting of route targets for all L2 virtual networks.
Enable L3 Gateway on Border - Select the check box to enable a Layer 3 gateway on the border switches.
A sample of the Create Network screen is given below.
-
Click Create Network. A message appears at the bottom right part of the screen indicating that the network is created.
The new network appears on the Networks page that comes up.
The Status is NA since the network is created but not yet deployed on the switches. Now that the network is created, you can create more networks if needed and deploy the networks on the devices in the fabric.
Export and Import Network Information
You can export network information to a .CSV file. The exported file contains information pertaining to each network, including the fabric it belongs to, the associated VRF, the network templates used to create the network, and all other configuration details that you saved during network creation.
In the Networks screen, click the Export icon to export network information as a .CSV file.
You can use the exported .CSV file for reference or use it as a template for creating new networks. To import networks, do the following:
-
Update new records in the .CSV file. Ensure that the networkTemplateConfig field contains the JSON Object. A message at the bottom right part of the screen displays errors and success messages. This screenshot depicts two new networks being imported.
-
In the Networks screen, click the Import icon and import the .CSV file into DCNM.
You can see that the imported networks are displayed in the Networks screen.
Editing Networks for the Standalone Fabric
Procedure
Step 1 |
Click Control > Networks. |
||
Step 2 |
Choose a fabric from the SCOPE drop-down list. |
||
Step 3 |
Choose a network. |
||
Step 4 |
Click the Edit icon. |
||
Step 5 |
Update the fields in the General and Advanced tabs of the Network Profile area as needed.
|
||
Step 6 |
Click Save at the bottom right part of the window to save the updates. |
Creating VRFs for the Standalone Fabric
-
Click Control > VRFs (under Fabrics submenu).
The VRFs screen comes up.
-
Choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists VRFs of the selected fabric.
-
Click the + button to add VRFs to the Standalone fabric. The Create VRF screen comes up. Most of the fields are autopopulated.
The fields in this screen are:
VRF ID and VRF Name: The ID and name of the VRF.
Note
For ease of use, the VRF creation option is also available while you create a network.
VRF Template: This template is applicable for VRF creation, and only applicable for leaf switches.
VRF Extension Template: The template is applicable when you extend the VRF to other fabrics, and is applicable for border devices.
Fill the fields in the VRF Profile section.
General tab – Enter the VLAN ID of the VLAN associated with the VRF, the corresponding Layer 3 virtual interface, and the VRF ID.
The VLAN ID default range is 2 to 3967. From DCNM Release 11.5(2), you can use a VLAN range greater than default value 3967. The reserved VLAN range must be set to a different range. In switch command enter “system vlan <vlan> reserve”. Save the configuration to startup configuration and reload the switch for the new reserved VLAN range to reflect.
From Cisco DCNM Web UI, choose Administration > DCNM Server > Server Properties, enter the value as 4094 for RM.TOP_DOWN_NETWORK_VLAN.MAX and RM.TOP_DOWN_VRF_VLAN.MAX, click Apply Changes and then restart DCNM. Once the DCNM is up, you can create VRF and network using the VLAN value greater than 3967.
Advanced tab – The fields in the tab are autopopulated.
VRF Intf MTU - Specifies VRF interface MTU.
Routing Tag – If a VLAN is associated with multiple subnets, then this tag is associated with the IP prefix of each subnet. Note that this routing tag is associated with overlay network creation too.
Redistribute Direct Route Map – Specifies the route map name for redistribution of routes in the VRF.
Max BGP Paths and Max iBGP Paths – Specifies the maximum BGP and iBGP paths.
TRM Enable – Select the check box to enable TRM.
If you enable TRM, then the RP address, and the underlay multicast address must be entered.
For more information, see Overview of Tenant Routed Multicast.
Is RP External – Enable this checkbox if the RP is external to the fabric. If this field is unchecked, RP is distributed in every VTEP.
RP Address – Specifies the IP address of the RP.
RP Loopback ID – Specifies the loopback ID of the RP, if Is RP External is not enabled.
Underlay Multicast Address – Specifies the multicast address associated with the VRF. The multicast address is used for transporting multicast traffic in the fabric underlay.Note
The multicast address in the Default MDT Address for TRM VRFs field in the fabric settings screen is auto-populated in this field. You can override this field if a different multicast group address should be used for this VRF.
Overlay Multicast Groups – Specifies the multicast group subnet for the specified RP. The value is the group range in “ip pim rp-address” command. If the field is empty, 224.0.0.0/24 is used as default.
Enable IPv6 link-local Option - Select the check box to enable the IPv6 link-local option under the VRF SVI. If this check box is unchecked, IPv6 forward is enabled.
Enable TRM BGW MSite - Select the check box to enable TRM on Border Gateway Multisite.
Advertise Host Routes – Enable the checkbox to control advertisement of /32 and /128 routes to Edge Routers.
Advertise Default Route – Enable the checkbox to control advertisement of default routes internally.
To allow inter-subnet communication between end hosts in different VXLAN fabrics, where the subnets are present in both fabrics, you must disable the Advertise Default Route feature (clear the Advertise Default Route checkbox) for the associated VRF. This will result in /32 routes for hosts being seen in both fabrics. For example, Host1 (VNI 30000, VRF 50001) in Fabric1 can send traffic to Host2 (VNI 30001, VRF 50001) in Fabric2 only if the host route is present in both fabrics. When a subnet is present in only one fabric then default route is sufficient for inter-subnet communication.
Config Static 0/0 Route - Select the check box to enable static default route configuration.
BGP Neighbor Password - Specifies the VRF Lite BGP neighbor password.
BGP Password Key Encryption Type - Select the encryption type from this drop-down list.
Sample screenshots of the Create VRF screen:
Advanced tab:
-
Click Create VRF.
The MyVRF_50001 VRF is created and appears on the VRFs page.
Export and Import VRF Information
You can export VRF information to a .CSV file. The exported file contains information pertaining to each VRF, including the fabric it belongs to, the templates used to create the VRF, and all other configuration details that you saved during VRF creation.
In the VRFs screen, click the Export icon to export VRF information as a .CSV file.
You can use the exported .CSV file for reference or use it as a template for creating new VRFs. To import VRFs, do the following:
-
Update new records in the .CSV file. Ensure that the vrfTemplateConfig field contains the JSON Object.
-
In the VRFs screen, click Import icon and import the .CSV file into DCNM.
A message at the bottom right part of the screen displays errors and success messages. This screenshot depicts a new VRF being imported.
Note
When you create a VRF using the Import option on the VRF window or using the DCNM APIs, you might see an error saying: Instance name is not specified.
This error is because of a tagging issue. To remove this error, edit the VRF in DCNM Web UI and then deploy.
You can see that the imported VRF is displayed in the VRFs screen.
Editing VRFs for the Standalone Fabric
-
Choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists VRFs of the selected fabric.
-
From the Select a Fabric drop-down list, select the fabric Standalone, and click Continue on the top right part of the screen. The Networks page is displayed.
-
Click the VRF View at the top right part of the screen. The VRFs page appears.
-
Select the VRF and click the Edit option at the top left part of the screen. The Edit VRF screen comes up.
-
Update the fields in the General and Advanced tabs of the VRF Profile section as needed.
-
Click Save at the bottom right part of the screen to save the updates.
Deploying Networks for the Standalone and MSD Fabrics
Before you begin: Ensure that you have created networks for the fabric.
-
Click Control > Networks (under Fabrics submenu).
The Networks screen comes up.
-
Choose the correct fabric from SCOPE. When you select a fabric, the Networks screen refreshes and lists networks of the selected fabric.
-
Select networks that you want to deploy. In this case, select the check boxes next to both the networks and click Continue at the top right part of the screen.
The Network Deployment page appears. On this page, you can see the network topology of the Standalone fabric.
You can deploy networks simultaneously on multiple switches. The selected devices should have the same role (Leaf, Border Gateway, and so on).
Note
In an MSD fabric, all member fabrics are visible from this screen.
At the bottom right part of the screen, the color codes that represent different stages of deployment are displayed. The color of the switch icons changes accordingly. Blue for Pending state, yellow for In Progress when the provisioning is in progress, green when successfully deployed, and so on. From DCNM 11.3(1), the pending state indicates that there is a pending deployment or pending recomputation. You can click on the switch and review the pending deployments using Preview or Deploy Config options, or click Save & Deploy to recompute the state of the switch.
The overlay networks (/VRFs) provisioning status is context-specific. It is a combination of networks that you chose for provisioning and the relevant switches in the topology. In this example, it means that the networks MyNetwork_30000 and MyNetwork_30001 are yet to be deployed on any switch in this fabric.
Undiscovered cloud display – To display (or not display) an Undiscovered cloud in this screen, click the cloud icon in the vertical panel, at the top-right part of the screen. When you click the icon, the Undiscovered cloud and its links to the fabric topology are not displayed. Click the icon again to display the Undiscovered cloud.
You can move the topology around the screen by clicking the left mouse button on the screen and moving it in the direction you desire. You can enlarge or shrink the switch icons proportionately by moving the cursor roller. You can also use corresponding alternatives on the touchpad.
-
Click ... in the Interfaces column.
The Interfaces box opens up. It lists interfaces or port channels. You can select interfaces/port channels to associate them with the selected network. For each interface, port type and description, channel number and connected neighbor interface details are displayed.
From Cisco DCNM Release 11.5(1), the Interfaces window doesn’t list interfaces that are part of an interface group. Specifically, the trunk ports, access ports, and dot1q tunnel ports.
If you try to perform a network attachment to a switch and an interface is part of an interface group, an appropriate error is displayed.
-
Double-click a switch to deploy the networks on it. For deployment of networks on multiple switches, click Multi-Select from the panel at the top right part of the screen (the topology freezes to a static state), and drag the cursor across the switches.
Immediately the Network Attachment dialog box appears.
A tab represents each network (the first network is displayed by default) that is being deployed. In each network tab, the switches are displayed. Each row represents a switch.
Click the check box next to the Switch column to select all switches. The network is ready to be provisioned on the switches.
VLAN - Update the VLAN ID if needed.
When you update a VLAN ID and complete the network deployment process, the old VLAN is not automatically removed. To complete the process, you should go to the fabric topology screen (click Control > Fabric Builder and click within the corresponding fabric box to go to the screen) and use the Save and Deploy option.
When updating the VLAN ID for a given network, the original VLAN ID is not automatically removed from the attached trunk interface. In order to remove the old or original VLAN ID, you must perform Save and Deploy + Config Deploy operation from within the fabric in Fabric Builder. For this, go to the fabric topology screen (click Control > Fabric Builder and click within the corresponding fabric box to go to the screen) and execute the Save and Deploy operation. Verify that config compliance is removing the expected config, then execute Deploy Config operation to remove the configs.
Interfaces – Click … in the column to add interfaces associated with the selected network.
VLAN to trunk port mapping – The selected trunk ports include the VLAN as an allowed VLAN on the port.
VLAN to vPC domain mapping - If you want to associate the VLAN to port channels of a vPC domain, add the port channels from the list of interfaces. The vPC port channels include the VLAN as an allowed VLAN.
Freeform configurations – Click Freeform config to enable additional configurations on the switch. After the configurations are saved, the Freeform config button gets highlighted.
-
Select the other network tab and make the same selections.
-
Click Save (at the bottom right part of your screen) to save the configurations.
Note
Addition and removal of interfaces are displayed in the Interfaces column of the Switches Deploy screen. Though the interface-related updates (like addition or removal of trunk ports) are provisioned on the switches, the correct configurations will not reflect in the preview screen. When you add or remove a trunk or access port, the preview shows the addition or removal of configurations for the interface under that network.
The topology window appears again. Click Refresh in the vertical panel at the top right part of the screen. The blue color on the switch icons indicates that the deployment is pending. From DCNM 11.3(1), the pending state indicates that there is a pending deployment or pending recomputation. You can click on the switch and review the pending deployments using Preview or Deploy Config options, or click Save & Deploy to recompute the state of the switch.
-
Preview the configurations by clicking Preview (the eye icon above the Multi-Select option). Since MyNetwork_30000 and MyNetwork_30001 are networks of VRF 50000, the configurations contain VRF configurations followed by the network configurations.
On the preview screen, you can select from the Select a switch and Select a network drop-down boxes at the top of the screen to view other network configurations.
After checking the configurations, close the screen. The Topology screen appears again.
-
Click Deploy on the top right part of the screen. The color of the switch icons changes to yellow and a message appears at the bottom right part of the screen indicating that the deployment is in progress. After the networks' deployment is complete, the color of the switch icons changes to green, indicating successful deployment.
Note
In case you click Deploy and there is no configuration diff that has to be deployed, a pop-up window comes up stating No switches PENDING for deployment.
Note
The status of the switch is determined by the aggregated status of the selected networks or VRFs in the following hierarchy: Pending, In Progress, Out-of-Sync/Failed, In Sync/Success, and Unknown/NA. For example, if any one of the networks or VRFs is in the Out-of-Sync/Failed status and others are not in the Pending or In Progress status, then the switch status is Out-of-Sync/Failed. The default status is Unknown/NA, when the status is not known.
Go to the Networks page to view the individual status for all networks.
Network Deployment for an MSD Fabric
Consider a scenario wherein you are deploying the same networks on different member fabric border devices. You can choose one fabric, deploy networks on its border devices, and then choose the second fabric and deploy networks.
Alternatively, you can choose the MSD fabric, and deploy the networks from a single topology view of all member fabric border devices.
This is a topology view of an MSD fabric wherein the two member fabrics topologies and their connections are depicted. You can deploy networks on the BGWs of the fabrics at once.
Detailed View
You can also use the Detailed View option to deploy networks and VRFs. Click Detailed View at the top right part of the screen. The Detailed View window appears. This lists the networks in a tabular view.
The options:
Edit - Select a network and click the Edit icon at the top left part of the screen.
Note |
If you select one network/switch entry and click on Edit, the Network Attach dialog box appears. To maintain consistency across the Topology View and Detailed View screens, the Network Attach screen displays all networks, and not just the selected network/switch. |
Preview – Click Preview to preview configurations before deployment. You can only preview pending configurations, and not uninitiated or deployed configurations.
Deploy – Click Deploy to provision networks onto the switches.
History – Select a row and click History to view the configuration instances and status. Network and VRF-wise configurations are displayed. Click in the Status column of any instance for more details.
The fields in the table contain the configuration instance in each row, the associated switch and fabric names, the switch role, trunk ports (if any), and the deployment status.
Quick Attach – Choose a network and click Quick Attach. A confirmation window appears. Click OK. The network will be attached to the selected switch.
Quick Detach– Choose a network and click Quick Detach. A confirmation window appears. Click OK. The network will be detached from the selected switch.
On the Detailed View page, the network profile configuration history is displayed. If you have associated specific trunk interfaces to that network, then the interface configuration is displayed as a separate configuration instance.
Note |
When you upgrade from an earlier release (such as DCNM 10.4[2]) to the DCNM 11.0(1) release, overlay networks and VRFs deployment history information from the earlier DCNM release is not retained. |
Deploying VRFs for the Standalone and MSD Fabrics
-
Choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists VRFs of the selected fabric.
-
Select check boxes next to the VRFs that you want to deploy and click Continue at the top right part of the screen.
The VRF Deployment screen appears. On this page, you can see the topology of the Standalone fabric. The following example shows you how to deploy the VRFs MyVRF_50000 and MyVRF_50001 on the leaf switch. You can deploy VRFs simultaneously on multiple switches but of the same role (Leaf, Border Gateway, and so on).
At the bottom right part of the screen, the color codes that represent different stages of deployment are displayed. The color of the switch icons changes accordingly. Blue for Pending state, yellow for In Progress state when the provisioning is in progress, red for failure state, green when successfully deployed, and so on. From DCNM 11.3(1), the pending state indicates that there is a pending deployment or pending recomputation. You can click on the switch and review the pending deployments using Preview or Deploy Config options, or click Save & Deploy to recompute the state of the switch.
The overlay networks (or VRFs) provisioning status is context-specific. It is a combination of VRFs that you chose for provisioning and the relevant switches in the topology. In this example, it means that the VRFs are yet to be deployed on any switch in this fabric.
Undiscovered cloud display – To display (or not display) an Undiscovered cloud in this screen, click the cloud icon in the vertical panel, at the top-right part of the screen. When you click the icon, the Undiscovered cloud and its links to the fabric topology are not displayed. Click the icon again to display the Undiscovered cloud.
You can move the topology around the screen by clicking the left mouse button on the screen and moving it in the direction you desire. You can enlarge or shrink the switch icons proportionately by moving the cursor roller. You can also use corresponding alternatives on the touchpad.
-
Double-click a switch to deploy VRFs on it. The VRF Attachment screen comes up.
Note
For deployment of VRFs on multiple switches, click the Multi-Select option from the panel at the top right part of the screen (This freezes the topology to a static state), and drag the cursor across the switches.
A tab represents each VRF that is being deployed (the first selected VRF is displayed by default). In each VRF tab, the selected switches are displayed. Each row represents a switch.
VLAN ID - Click within the VLAN column to update the VRF VLAN ID, if needed.
Freeform configurations – Click Freeform config to enable additional configurations on the switch. After you save freeform configurations, the Freeform config button gets highlighted.
Click the checkbox next to the Switch column to select all switches. VRF MyVRF_50000 is ready to be provisioned on the switch
-
Select the other VRF tab and make the same selections.
-
Click Save (at the bottom right part of your screen) to save VRF configurations.
The topology screen comes up again. Click the Refresh button in the vertical panel at the top right part of the screen. The blue color on the switch icons indicates that the deployment is pending. From DCNM 11.3(1), the pending state indicates that there is a pending deployment or pending recomputation. You can click on the switch and review the pending deployments using Preview or Deploy Config options, or click Save & Deploy to recompute the state of the switch.
Preview the configurations by clicking the Preview button (the eye icon above the Multi-Select option).
After checking the configurations, close the screen. The Topology View screen appears.
-
Click the Deploy button on the top right part of the screen. The color of the switch icons changes to yellow and a message appears at the bottom right part of the screen indicating that the deployment is in progress. After the VRF deployment is complete, the color of the switch icons changes to green, indicating successful deployment.
Note
In case you click Deploy and there is no configuration diff that has to be deployed, a pop-up window comes up stating No switches PENDING for deployment.
VRFs Deployment for an MSD Fabric
Consider a scenario wherein you are deploying the same VRFs on different member fabric border devices. You can choose one fabric, deploy VRFs on its border devices, and then choose the second fabric and deploy the VRFs.
Alternatively, you can choose the MSD fabric, and deploy the VRFs from a single topology view of all member fabric border devices at once.
Detailed View
You can also use the Detailed View button to deploy networks and VRFs.
Click Detailed View at the top right part of the screen. The Detailed View screen comes up. This lists the VRFs in a tabular view.
The options:
Edit - Select a VRF and click the Edit icon at the top left part of the screen.
Note |
If you select one VRF/switch entry, the VRF Attach screen comes up. To maintain consistency across the Topology View and Detailed View screens, the VRF Attach screen displays all VRFs, and not just the selected VRF/switch entry. |
Preview – Click Preview to preview configurations before deployment. You can only preview pending configurations, and not uninitiated or deployed configurations.
Deploy – Click Deploy to provision VRFs onto the switches.
History – Select a row and click History to view the configuration instances and status. Network and VRF-wise configurations are displayed. Click in the Status column of any instance for more details.
The fields in the table contain the configuration instance in each row, the associated switch and fabric names, the switch role, and the deployment status.
Quick Attach: Choose a VRF and click Quick Attach. A confirmation window appears. Click OK. The VRF will be attached to the selected switch.
Quick Detach: Choose a VRF and click Quick Detach. A confirmation window appears. Click OK. The VRF will be detached from the selected switch.
Note |
When you upgrade from an earlier release (such as DCNM 10.4[2]) to the DCNM 11.0(1) release, overlay networks and VRFs deployment history information from the earlier DCNM release is not retained. |
Undeploying Networks for the Standalone Fabric
You can undeploy VRFs and networks from the deployment screen. The DCNM screen flow for undeployment is similar to the deployment process flow. Go to the deployment screen (Topology View) to undeploy networks:
-
Click Control > Networks (under Fabrics submenu).
The Networks screen comes up.
-
Choose the correct fabric from SCOPE. When you select a fabric, the Networks screen refreshes and lists networks of the selected fabric.
-
Select the networks that you want to undeploy and click Continue. The topology view comes up.
-
Select the Multi-Select button (if you are undeploying the networks from multiple switches), and drag the cursor across switches with the same role. The Network Attachment screen comes up.
(For a single switch, double-click the switch and the Network Attachment screen comes up).
(For a single switch, double-click the switch and the Switches Deploy screen comes up).
-
In the Network Attachment screen, the Status column for the deployed networks is displayed as DEPLOYED. Clear the check boxes next to the switches, as needed. Ensure that you repeat this on all tabs since each tab represents a network.
-
Click Save (at the bottom right part of the screen) to initiate the undeployment of the networks. The Topology View comes up again.
Note
Alternatively, you can click the Detailed View button to undeploy networks.
-
Refresh the screen, preview configurations if needed and click Deploy to remove the network configurations on the switches. After the switch icons turn green, it indicates successful undeployment.
-
Go to the Networks page to verify if the networks are undeployed.
Undeploying VRFs for the Standalone Fabric
You can undeploy VRFs from the deployment screen. The DCNM screen flow for undeployment is similar to the deployment process flow.
-
Choose Control > Fabrics > VRFs.
-
Choose the correct fabric from SCOPE. When you select a fabric, the VRFs screen refreshes and lists networks of the selected fabric.
-
Select the VRFs that you want to undeploy and click Continue. The Topology View page comes up.
-
Select the Multi-Select option (if you are undeploying the VRFs from multiple switches), and drag the cursor across switches with the same role. The VRF Attachment screen comes up.
(For a single switch, double-click the switch and the VRF Attachment screen comes up).
-
In the Switches Deploy screen, the Status column for the deployed VRFs is displayed as DEPLOYED. Clear the check boxes next to the switches, as needed. Ensure that you repeat this on all tabs since each tab represents a VRF.
-
Click Save (at the bottom right part of the screen) to initiate the undeployment of the VRFs. The topology view comes up again.
Note
Alternatively, you can click the Detailed View button to undeploy VRFs.
-
Refresh the screen, preview configurations if needed and click Deploy to remove the VRF configurations on the switches. After the switch icons turn green, it indicates successful undeployment.
-
Go to the VRFs page to verify if the networks are undeployed.
Deleting Networks and VRFs
If you want to delete networks and corresponding VRFs in the MSD fabric, follow this order:
-
Undeploy the networks, if not already done.
-
Delete the networks.
-
Undeploy the VRFs, if not already done.
-
Delete the VRFs.
Configuring Multiple VLAN IDs to a Single VNI
The following procedure shows how to tag multiple VLAN IDs to a single VNI in DCNM.
Procedure
Step 1 |
Navigate to Control > Networks. |
Step 2 |
Select the fabric from the Scope drop-down list and then select the network. Click Continue. |
Step 3 |
Check the Multi-Select check box and drag the cursor over the switches that needs to be updated with VLAN IDs. |
Step 4 |
In the Network Attachment window, edit the VLAN ID for the switches and click Save. |
Step 5 |
Click Deploy to deploy the configuration. |
Enhanced Role-based Access Control in Cisco DCNM
From Cisco DCNM Release 11.4(1), you can see the following role-based access control (RBAC) changes:
-
Read-only access to the Cisco DCNM Web UI and APIs for the network-operator user role
-
A new user role called network-stager.
-
Freeze deployment for a particular fabric or all fabrics in DCNM as a user with the network-admin role.
From Cisco DCNM Release 11.5(1), you can see new user roles, device-upg-admin, and access-admin are added.
Note |
Actions that cannot be performed by a selected user role is grayed out. |
You can also watch the video that demonstrates some of the operations performed by a network stager and how to freeze a fabric in Cisco DCNM. See the Enhanced Role-based Access Control (RBAC) video.
Device-upg-admin Role
A user with the device-upg-admin role can perform operations only in Image Management window.
See the Image Management section for more information.
Access-admin Role
A user with the access-admin role can perform operations only in Interface Manager window for all fabrics.
An access-admin can perform the following actions:
-
Add, edit, delete and deploy layer 2 port channels, and vPC.
-
Edit host vPC, and ethernet interfaces.
-
Save, preview, and deploy from management interfaces.
-
Edit interfaces for LAN classic fabrics.
Apart from nve, management, tunnel, subinterface, SVI, interface grouping, and loopback interfaces
However, a user with the access-admin role can’t perform the following actions:
-
Cannot edit layer 3 port channels, ST FEX, AA FEX, loopback interfaces, nve interfaces, and subinterfaces.
-
Cannot edit member interfaces and port channels of Layer 3, ST FEX, AA FEX.
-
Cannot edit interfaces with policy associated from underlay and link for easy fabrics.
-
Cannot edit peer link port channel.
-
Cannot edit management interface.
-
Cannot edit tunnel.
Note |
The icons and buttons are grayed out for this role when the fabric or DCNM is in deployment-freeze mode. |
Network-Operator Role
A user with the network-operator role has access to the following menu in the DCNM Web UI:
-
Dashboard
-
Topology
-
Monitor
-
Applications
From Cisco DCNM, Release 11.4(1), a user with this role has read-only access to the Control menu as well.
A network operator can view fabric builder, fabric settings, preview configurations, policies, and templates. However, a network operator cannot perform the following actions:
-
Cannot change expected configurations of any switch within any fabric.
-
Cannot deploy any configurations to switches.
-
Cannot access the administration options like licensing, creating more users, and so on.
Network-Stager Role
A user with the network-stager role can make configuration changes on DCNM. A user with the network-admin role can deploy these changes later. A network stager can perform the following actions:
-
Edit interface configurations.
-
View or edit policies.
-
Create interfaces.
-
Change fabric settings.
-
Edit or create templates.
However, a network stager cannot perform the following actions:
-
Cannot make any configuration deployments to switches.
-
Cannot perform deployment-related actions from the DCNM Web UI or the REST APIs.
-
Cannot access the administration options like licensing, creating more users, and so on.
-
Cannot move switches in and out of maintenance mode.
-
Cannot move fabrics in and out of deployment-freeze mode.
-
Cannot install patches.
-
Cannot upgrade switches.
-
Cannot create or delete fabrics.
-
Cannot import or delete switches.
The difference between a network operator and a network stager is as a network stager, you can only define intent for existing fabrics, but cannot deploy those configurations.
Only a network admin can deploy the changes and edits that are staged by a user with the network-stager role.
Viewing Policy Change History
Different users can simultaneously change expected configuration of switches in the DCNM. You can view the history of these staged changes in the Policy Change History tab. The deployment history captures the changes that are pushed or deployed from DCNM to switches.
Note |
Only deployment history is supported for non-Nexus devices. |
To view the changes by different users, perform the following steps:
Procedure
Step 1 |
Log into Cisco DCNM with the network-admin, network-stager, or network-operator user role. |
Step 2 |
Navigate to the fabric topology window. |
Step 3 |
Right-click the switch for which you intent change history. |
Step 4 |
Choose History. |
Step 5 |
Click the Policy Change History tab. |
Step 6 |
Search for the interface to which you made changes in the Generated Config column. |
Step 7 |
The PTI Operation column will have the value UPDATE for the changes made by different users. |
Step 8 |
Scroll horizontally to the User column. You can see the user names populated with the timestamp. For each configurable entity, the detailed history under the Generated Config column, provides delta of configuration changes made by every user. |
Freezing Fabrics in Cisco DCNM
As a network admin, you can disable, or freeze, deployments for LAN classic fabrics, easy fabrics, and external fabrics. Deployment freeze disables configuration or write access from the DCNM to the switches. When you freeze a fabric, switches cannot be reloaded, moved in and out of maintenance mode, and you cannot add or delete switches within the fabric. This feature provides complete control for a network admin to disable inadvertent changes to the physical network from the DCNM, unless a maintenance window is scheduled.
Freezing a Fabric
To disable deployment for a fabric from Cisco DCNM Web UI, perform the following:
Procedure
Step 1 |
Navigate to the Fabric Builder window or the fabric topology window. |
||
Step 2 |
Click the spanner () icon. The spanner icon is next to the fabric name in the fabric topology window. A confirmation window appears asking you if you want to disable all deployments for the fabric. |
||
Step 3 |
Click Yes.
After you disable deployments or freeze a fabric, you can only save, edit, or preview changes but not deploy them. All deploy related actions from the DCNM to this fabric will be grayed out. To enable all deployments for the fabric, click the same spanner () icon and unfreeze the fabric. |
Freezing All Fabrics
In addition to the per-fabric deployment freeze knob, the network admin can freeze deployments for all fabrics within the DCNM at the same time.
To freeze all fabrics in your DCNM setup from the Cisco DCNM Web UI, perform the following steps:
Procedure
Step 1 |
Choose Administration > DCNM Server > Server Properties. |
||
Step 2 |
Search for the DEPLOYMENT_FREEZE field. |
||
Step 3 |
Set the value as true. The default value is false.
Actions that cannot be performed when you freeze a fabric or DCNM are grayed out. |
Fabric Backup and Restore
This section describes the fabric backup and restore in Cisco DCNM.
Backing Up Fabrics
You can back up all fabric configurations and intents automatically or manually. You can save configurations in DCNM, which are the intents. The intent may or may not be pushed on to the switches.
DCNM doesn’t back up the following fabrics:
-
External fabrics in monitor-only mode: Backing up of external fabrics in monitor-only mode isn’t supported because you can’t restore any configurations or intent. However, if such external fabrics are member fabrics of an MSD fabric, the backup is taken at MSD-fabric level.
Note
From Cisco, DCNM Release 11.4(1), you can take a backup of external fabrics in monitor-only mode, but can’t restore them. You can restore this backup when the external fabric isn’t in monitor-only mode.
-
Parent MSD fabrics in releases earlier than Cisco DCNM, Release 11.4(1): You can only back up the configurations and intent of member fabrics in an MSD fabric individually.
Note
From Cisco DCNM, Release 11.4(1), you can take backups of MSD fabrics. When you initiate a backup from the parent fabric, the backup process is applicable for the member fabrics as well. However, DCNM stores all the backed-up information of the member fabrics and the MSD fabric together in a single directory.
From Cisco DCNM Release 11.4(1), the backup captures the intent related to IFC as well. When you’re backing up an external fabric, the checkpoints are copied from the switches to DCNM. The backup configuration files are stored in the following path in DCNM: /usr/local/cisco/dcm/dcnm/data/archive
The backed-up config files can be found in the
corresponding directory with the fabric name. Each backup of a fabric is treated as
a different version, regardless if it is backed up manually or automatically. You
can find all versions of the backup in the corresponding fabric directories. Hence,
the backed up intent configuration file, running configuration file and PTIs can be
found at location:
/usr/local/cisco/dcm/dcnm/data/archive/<fabric_name>/Version_x
,
where x
is the version number. The valid value is between 1 and the
limit you set in the archived.versions.limit field. The default value is 50,
which means only 50 backups are archived, and the oldest backups are removed. The
minimum value is 10. If you specify a value lesser than 10, it will be overwritten
to 10.
You can set the
number of backup files to be archived in the Server Properties window. Search
for the # Number of archived files per fabric to be retained: section in the
Server Properties window. Enter a value in the
archived.versions.limit field.
You can also watch the video that demonstrates how to back up and restore an MSD fabric in Cisco DCNM. See the MSD Fabric Backup and Restore video.
Backing Up Fabrics Automatically
You can enable an automatic hourly backup or scheduled backup for fabric configurations and intents. There are two types of automatic backup.
The backup has the information related to intent and fabric configurations in addition to associated state of the resource manager in terms of used resources on fabrics. DCNM backs up only when there’s a configuration push. DCNM triggers the automatic backup only if you didn’t trigger any manual backup after the last configuration push.
There are two types of automatic backup.
-
Hourly Fabric Backup: You can enable an hourly backup.
Note
MSD fabrics don’t support hourly backup.
-
Scheduled Fabric Backup: You can schedule a fabric backup for regular intervals.
Note |
In external fabrics, DCNM backs up the changes in the running configurations as well. The configuration push happens after a deploy. If you didn’t deploy the changes, you can’t back up them in an intent. |
Hourly and scheduled backup processes happen only during the next periodic configuration compliance activity, and there can be a delay of up to an hour.
Hourly and Scheduled Backup of Fabrics
To enable automatic backup of fabric configurations and intents from the Cisco DCNM Web client, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder. The Fabric Builder window appears. |
||
Step 2 |
Click the Edit Fabric icon for the fabric you want to backup. |
||
Step 3 |
Click the Configuration Backup tab. |
||
Step 4 |
Choose the nature of backup by checking the appropriate check box. The valid options are Hourly Fabric Backup and Scheduled Fabric Backup. If you want to enable both the backups, check the Hourly Fabric Backup check box and the Scheduled Fabric Backup check box.
|
||
Step 5 |
Click Save. DCNM initiates the backup process after you click Save. |
Backing Up Fabrics Manually
You can enable a manual backup for fabric configurations and intents. Regardless of the settings you choose under the Configuration Backup tab in the Edit Fabric dialog box, you can initiate a backup using this option. You cannot initiate standalone backups for a member fabric on an MSD fabric.
To initiate a manual backup of fabric configurations and intents from the Cisco DCNM Web UI, perform the following steps:
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder. The Fabric Builder window appears. |
||
Step 2 |
Click the fabric for which you want to backup immediately. The fabric topology window appears. |
||
Step 3 |
Click Backup Now in the Actions pane. The Backup Now dialog appears. |
||
Step 4 |
Enter a tag name in the Tag field. |
||
Step 5 |
Click OK. A confirmation message appears that the backup is triggered successfully.
|
||
Step 6 |
(Optional) Click Restore Fabric from the Actions pane to confirm if the manual backup is successful or not. The manual backup is indicated in midnight blue. When you hover over the backup, the name has the tag you mentioned in Step 4 confirming that it’s a manual backup. |
Golden Backup
You can now mark the backups that you don’t want to delete even after you reach the archiving limit. These backups are the golden backups. You can’t delete golden backups of fabrics. However, Cisco DCNM archives only up to 10 golden backups. You can mark a backup as golden backup while restoring the fabric. To mark a backup as golden in Cisco DCNM, perform the following steps from the Cisco DCNM Web UI:
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder and select a fabric. |
Step 2 |
Click Restore Fabric from the Actions menu. The Restore Fabric window appears. |
Step 3 |
Choose the time period from where you want to choose the backup. Valid values are 1m, 3m, 6m, YTD, 1y, and All. You can zoom into the graph. By default the backup information for 1m, which is one month, appears. You can also choose a custom date range. The backup information includes the following information:
|
Step 4 |
Choose the backup you want to mark as golden by clicking the backup. You can choose the automatic or manual backup. These backups are color-coded. Automatic backups are indicated in blue color. Manual backups are indicated in midnight blue color. Golden backups are indicated in orange color. The automatic backups have only the versions in their names. Whereas the manual backups have tag names, which you gave when you initiated a manual backup, along with the version in the backup name. Hover over a backup to see the name. The automatic backup is initiated from the Configuration Backup tab in the Fabric Settings dialog box. The manual backup is initiated by clicking Backup Now from the Actions pane in the fabric topology window. |
Step 5 |
Check the Mark backup as golden backup check box to mark the backup as a golden backup. A confirmation dialog box appears. |
Step 6 |
Click Yes. |
Step 7 |
Continue with rest of the fabric restore procedure as mentioned in the Restoring Fabrics section or exit the window. |
Validating Backups
DCNM validates all the backups when you initiate a fabric restore process. The validation includes the following checks:
-
The DCNM release from which you want to restore: You can restore backups only from Cisco DCNM, Release 11.3(1), and Cisco DCNM, Release 11.4(1). Hence, if you upgrade from Cisco DCNM, Release 11.3(1) to Cisco DCNM, Release 11.4(1), you can restore a backup, which you archived before upgrading.
-
Member fabrics composition: DCNM checks the name or ID of the member fabrics of an MSD fabric. If you change them after you back up, the restore won’t proceed.
-
Template validation: DCNM checks if templates from the backup match the templates in the current version. If you delete or rename any templates, you can’t proceed with the restoring.
-
Device composition of a fabric: If there are any changes to the inventory of switches after you back them up, you can’t restore.
Restoring Fabrics
This section describes the fabric restoring for different types of fabrics. Cisco DCNM supports configuration restore at fabric level. Take a backup of the configuration to restore it.
Note |
After a backup and restore operation, the capabilities set to
Then restart the services. |
Restoring Easy Fabrics
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder and select a fabric. |
||
Step 2 |
Select Restore Fabric from the Actions menu. The Restore Fabric window appears. |
||
Step 3 |
Choose the time for which you want to restore the configuration. Valid values are 1m, 3m, 6m, YTD, 1y, and All. You can zoom into the graph. By default the backup information for 1m, which is one month, appears. You can also select a custom date range. The backup information includes the following information:
|
||
Step 4 |
Choose the backup you want to restore. You can choose the automatic or manual backup. These backups are color-coded. Automatic backups are indicated in blue color. Manual backups are indicated in midnight blue color. Golden backups are indicated in orange color. The automatic backups have only the versions in their names. Whereas the manual backups have tag names, which you gave when you initiated a manual backup, along with the version in the backup name. Hover over a backup to see the name. The automatic backup is initiated from the Configuration Backup tab in the Fabric Settings dialog box. The manual backup is initiated by clicking Backup Now from the Actions pane in the fabric topology window.
|
||
Step 5 |
Check the Mark backup as golden backup check box to mark the backup as a golden backup. |
||
Step 6 |
Click Next to see the selected backup information of the devices in sync. The switch name, switch serial number, IP address, and the delta configuration details of the devices appear.
|
||
Step 7 |
Click Get Config to preview the configuration details. Config Preview window appears, which has two tabs.
|
||
Step 8 |
Go back to View Backup Summary window. |
||
Step 9 |
Click Restore Intent to proceed with the restoring. The Restore Status window appears. You can view the status of the following:
The valid values for the status of any action are In Progress, Pending, or Failed.
|
||
Step 10 |
Click Next after the intent is restored. The Configuration Preview window appears. You can view the following details in this window:
|
||
Step 11 |
Click Deploy to deploy the restored configuration. The Configuration Deployment Status window appears. You can view the details of the switch name, IP address, status, status description, and the progress. |
||
Step 12 |
Click Close after the restoring process is complete. |
Restoring External Fabrics
When you restore an external fabric, the backed-up checkpoint is copied from DCNM to switches. To restore an external fabric in Cisco DCNM, perform the following steps from the Cisco DCNM Web UI:
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder and select a fabric. |
||||
Step 2 |
Select Restore Fabric from the Actions menu. The Restore Fabric window appears. |
||||
Step 3 |
Select the time for which you want to restore the configuration. Valid values are 1m, 3m, 6m, YTD, 1y, and All. You can zoom into the graph. By default the backup information for 1m, which is one month, appears. When you select a backup version, the vertical bar representing it turns grey, and corresponding information is displayed at the bottom part of the screen. It includes the following information:
You can select a custom date range either by rearranging the date slide below the vertical bars, or using the From and To boxes at the top right part of the screen. |
||||
Step 4 |
Choose the backup you want to restore. You can choose the automatic or manual backup. These backups are color-coded. Automatic backups are indicated in blue color. Manual backups are indicated in midnight blue color. Golden backups are indicated in orange color. The automatic backups have only the versions in their names. Whereas the manual backups have tag names, which you gave when you initiated a manual backup, along with the version in the backup name. Hover over a backup to see the name. The automatic backup is initiated from the Configuration Backup tab in the Fabric Settings dialog box. The manual backup is initiated by clicking Backup Now from the Actions pane in the fabric topology window.
|
||||
Step 5 |
(Optional) Check the Mark backup as golden backup check box to mark the backup as a golden backup. |
||||
Step 6 |
Click Next to see the selected backup information of the devices in sync. The switch name, switch serial number, IP address, status, Restore Supported (indicating whether the device supports checkpoint rollback or not), the configuration details of the devices, and the VRF appear.
By default, the management VRF is displayed in the VRF column because it is used for the copy operation during the restore process. If you want to use a different VRF for the copy operation, update the VRF column. To update the same VRF for all devices, use the Apply for all devices option at the bottom-left part of the screen. A sample screenshot:
|
||||
Step 7 |
Click Get Config to preview device configuration details. The Config Preview window appears, which has three tabs.
|
||||
Step 8 |
Go back to the View Backup Summary window. |
||||
Step 9 |
Click Restore Intent to proceed with the restoring. The Restore Status window appears. You can view the status of the following:
The valid values for the status of any action are In Progress, Pending, Completed, or Failed.
|
||||
Step 10 |
Click Close after the restore process is complete. |
Restoring MSD Fabrics
When you restore an MSD fabric, the overlay information related to the MSD fabric is restored before restoring information related to the child fabrics. If there’s any change in the inventory of the MSD fabric, the backup is considered to be invalid and the restore is blocked. You can’t initiate a restore process for a member fabric. You get an error stating that the fabric is currently a member fabric of an MSD fabric. Move the member fabrics out of the MSD fabric to restore the previous standalone backups. Restoring an MSD fabric involves restoring fabric intent, underlay or interface intent, overlay intent, and intent regeneration.
You can also watch the video that demonstrates how to backup and restore an MSD fabric in Cisco DCNM. See the MSD Fabric Backup and Restore video.
To restore an easy fabric in Cisco DCNM, perform the following steps from the Cisco DCNM Web UI:
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder. |
||
Step 2 |
Choose an MSD fabric. |
||
Step 3 |
Click Restore Fabric from the Actions menu. The Restore Fabric wizard appears and you will be in the Select Backup step.
|
||
Step 4 |
Choose the time for which you want to restore the configuration. Valid values are 1m, 3m, 6m, YTD, 1y, and All. You can zoom into the graph. By default the backup information for 1m, which is one month, appears. You can also select a custom date range. The backup information includes the following information:
|
||
Step 5 |
Choose the backup you want to restore. You can choose the automatic or manual backup. These backups are color-coded. Automatic backups are indicated in blue color. Manual backups are indicated in midnight blue color. Golden backups are indicated in orange color. The automatic backups have only the versions in their names. Whereas the manual backups have tag names, which you gave when you initiated a manual backup, along with the version in the backup name. Hover over a backup to see the name. The automatic backup is initiated from the Configuration Backup tab in the Fabric Settings dialog box. The manual backup is initiated by clicking Backup Now from the Actions pane in the fabric topology window.
|
||
Step 6 |
Click the backup you want to restore. The Backup Summary area appears. It includes the following information:
|
||
Step 7 |
Click Next to move to the Restore Preview step. The Easy Fabric tab has information about the switch name, fabric name, switch serial, IP address, and the delta configuration of the member easy fabric. The External Fabric tab contains information about the switch name, fabric name, switch serial, IP address, switch status, configuration, and if the restore is supported for the member external fabric.
|
||
Step 8 |
Click Restore Intent to proceed to the Restore Status step in restoring. The restore status and description appears for the member fabrics. Click the member fabric radio button to view the fabric-level progress of that fabric. The progress is automatically updated every 5 seconds. |
||
Step 9 |
Click Next after the status is successful. The Configuration Preview window appears. You can view the details of the switch name, IP address, switch serial number, preview configuration, status, and the progress in this window.
|
||
Step 10 |
Click Deploy to deploy the restored configuration. The Configuration Deployment Status window appears. You can view the following details:
|
||
Step 11 |
Click Close after the restoring process is complete. |
Restoring a Switch
From Cisco DCNM, Release 11.5(1), you can restore a Cisco Nexus switch in external fabrics and LAN classic fabrics from the Cisco DCNM Web UI. The information you restore at switch-level is extracted from the fabric-level backups. The switch-level restoring doesn’t restore fabric-level intents and other configurations applied using the fabric settings. Only switch-level intents are restored. Therefore, after you restore a switch, it might go out-of-sync because the fabric-level intents aren’t restored. Perform a fabric-level restore to restore the intents as well. You can restore only one switch at a time. You can’t restore a switch if the fabric where it’s discovered is part of an MSD fabric.
To restore a switch in Cisco DCNM, perform the following steps from the Cisco DCNM Web UI:
Procedure
Step 1 |
Choose Control > Fabrics > Fabric Builder. |
||
Step 2 |
Choose an external fabric or LAN classic fabric. |
||
Step 3 |
Right-click a Cisco Nexus switch for which you want to restore the configurations. |
||
Step 4 |
Choose the Restore Config option. Alternatively, you can click Tabular view in the Actions pane and navigate to the Switches tab. Choose a Cisco Nexus switch by checking the check box and click Restore. For non-Nexus switches, the Restore Config option doesn’t appear and the Restore button grays out. This option does not appear when you log in with the network-operator role or when the fabric is in monitor mode or freeze mode.The Restore Switch wizard appears and you are in the Select Backup step. |
||
Step 5 |
Choose the time for which you want to restore the configuration. Valid values are 1m, 3m, 6m, YTD, 1y, and All. You can zoom into the graph. By default the backup information for 1m, which is one month, appears. You can also select a custom date range. |
||
Step 6 |
Choose the backup you want to restore. You can choose the automatic, manual, or golden backup. These backups are color-coded. Automatic backups are indicated in blue color. Manual backups are indicated in midnight blue color. Golden backups are indicated in orange color. The automatic backups have only the versions in their names. Whereas the manual backups have tag names, which you gave when you initiated a manual backup, along with the version in the backup name. Hover over a backup to see the name. The automatic backup is initiated from the Configuration Backup tab in the Fabric Settings dialog box. The manual backup is initiated by clicking Backup Now from the Actions pane in the fabric topology window. |
||
Step 7 |
Click the backup you want to restore. The Backup Summary area appears. It includes the following information:
|
||
Step 8 |
Click Next to move to the Restore Preview step. You can view information about the switch name, switch serial, IP address, status, restore supported, delta configuration and the VRF details. |
||
Step 9 |
(Optional) Click Get Config to preview device configuration details. The Config Preview window appears, which has three tabs.
|
||
Step 10 |
Click Restore to proceed to the Restore Status step in restoring. The restore status and description appears for the switch. |
||
Step 11 |
Click Close after the restoring process is complete.
|
Deleting a VXLAN BGP EVPN Fabric
Choose Control > Fabric Builder. On the Fabric Builder page, click X on the rectangular box that represents the fabric. Ensure the following before deleting a fabric.
-
Fabric devices should not be in transition such as migration into or out of the fabric, ongoing network or VRF provisioning, and so on. Delete a fabric after the transition is complete.
-
Remove devices that are still attached to the fabric. Remove non-Cisco Nexus 9000 Series switches first and then remove the 9000 Series switches.
Post DCNM 11.5(1) Upgrade for VXLAN BGP EVPN, External, and MSD Fabrics
Note the following guidelines after you upgrade to the DCNM Release 11.5(1):
-
As part of the upgrade from an earlier DCNM release, the fabric and associated templates are carried over to the DCNM Release 11.5(1).
-
From DCNM 11.3(1), some of the policy templates from earlier DCNM releases are deprecated and are actively updated in each newer DCNM release. These policy templates are automatically removed after the upgrade if they are not found to be in use. This removal does not affect any operations and helps in reducing the number of policies displayed in the DCNM template library.
-
Navigate to each fabric from the Fabric Builder window, and click Save & Deploy to deploy any changes.
If you encounter any new or unexpected pending configurations after you click Save & Deploy, refer Configuration Compliance in DCNM.
Caution
Some configuration changes can be expected as part of this step. Therefore, perform it only during a scheduled maintenance window.
-
Post DCNM upgrade from Release 11.2(1), you could see the following diff if the fabric has a border device (border, border spine, border gateway, etc):
route-map extcon-rmap-filter-v6 deny 20 no match ip address prefix-list host-route-v6 route-map extcon-rmap-filter-v6 deny 20 match ipv6 address prefix-list host-route-v6
The above config is expected and it is meant to correct the route-map definition. Deployment of this diff will correct the switch configuration. If the fabric was created as Greenfield before upgrade, no additional action is needed. If the fabric was created as Brownfield before upgrade with the wrong route-map configuration on the device, this config will be captured in a switch_freeform policy. Post upgrade, you should edit the freeform policy to remove the CLI match ip address prefix-list host-route-v6 before the deployment.
-
After a multi-level upgrade from Cisco DCNM 10.4(2) or 11.0(1), you can change the VRF templates to Default_VRF_Universal or Default_VRF_Extension_Universal to enable ipv6 address use-link-local-only.
Changing ISIS Configuration from Level 1 to Level 2
This procedure shows how to change ISIS configuration on switches from Level 1 to Level 2 in a VXLAN fabric deployment.
-
Choose Control > Fabrics > Fabric Builder.
-
Click a fabric in the Fabric Builder window.
-
Click Tabular view under Actions menu.
-
Search for all the base_isis policies in the Template search field.
-
Select all the base_isis policies and click the Delete icon to delete policies
-
Click Save & Deploy.
After all the base_isis policies are deleted, DCNM considers the migrated brownfield fabric as a greenfield fabric and creates the base_isis_level2 policies on the switches.
Configuration Compliance in DCNM
The entire intent or expected configuration defined for a given switch is stored in DCNM. When you want to push this configuration down to one or more switches, the configuration compliance (CC) module is triggered. CC takes the current intent, the current running configuration, and then comes up with the set of configurations that are required to go from the current running configuration to the current expected config so that everything will be In-Sync.
When performing a software or firmware upgrade on the switches, the current running configuration on the switches is not changed. Post upgrade, if CC finds that the current running configuration does not have the current expected configuration or intent, it reports an Out-of-Sync status. There is no auto deployment of any configurations. You can preview the diffs that will get deployed to get one or more devices back In-Sync.
With CC, the sync is always from the DCNM to the switches. There is no reverse sync. So, if you make a change out-of-band on the switches that conflicts with the defined intent in DCNM, CC captures this diff, and indicates that the device is Out-of-Sync. The pending diffs will undo the configs done out-of-band to bring back the device In-Sync. Note that such conflicts due to out-of-band changes are captured by the periodic CC run that occurs every 60 mins by default, or when you click the RESYNC option either on a per fabric or per switch basis. Note that you can also capture the out-of-band changes for the entire switch by using the CC REST API. For more information, see Cisco DCNM REST API Guide, Release 11.2(1).
From Cisco DCNM Release 11.2(1), to improve ease of use and readability of deployed configurations, CC in DCNM has been enhanced with the following:
-
All displayed configurations in DCNM are easily readable and understandable.
-
Repeated configuration snippets are not displayed.
-
Pending configurations precisely show only the diff configuration.
-
Side-by-side diffs has greater readability, integrated search or copy, and diff summary functions.
The CC engine computes diff by comparing the intent with the running configuration on the switch ensuring that any configuration that is defined in the intent exists on the switch. For any component or configuration snippet that is defined in the intent, the CC engine ensures that the same component or configuration snippet exists on the switch by generating appropriate commands, if required, to match the switch configuration with the intent configuration.
Top-level configuration commands on the switch that do not have any associated DCNM intent are not checked for compliance by Configuration Compliance (CC). However, CC performs compliance checks, and attempts removal, of the following commands even if there is no DCNM intent:
-
configure profile
-
apply profile
-
interface vlan
-
interface loopback
-
interface Portchannel
-
Sub-interfaces, for example, interface Ethernet X/Y.Z
-
fex
-
vlan <vlan-ids>
CC performs compliance checks, and attempts removal, of these commands only when Easy_Fabric_11_1 and Easy_Fabric_eBGP fabric templates are used. On External_Fabric templates, top-level configuration commands on the switch, including the commands mentioned above, that do not have any associated DCNM intent are not checked for compliance by CC.
We recommend using the DCNM freeform configuration template to create additional intent and deploy these commands to the switches to avoid unexpected behavior
Now, consider a scenario in which the configuration that exists on the switch has no relationship with the configuration defined in the intent. Examples of such configurations are a new feature that has not been captured in the intent but is present on the switch or some other configuration aspect that has not been captured in the intent. Configuration compliance does not consider these configuration mismatches as a diff. In such cases, Strict Configuration Compliance ensures that every configuration line that is defined in the intent is the only configuration that exists on the switch. However, configuration such as boot string, rommon configuration, and other default configurations are ignored during strict CC checks. For such cases, the internal configuration compliance engine ensures that these config changes are not called out as diffs. These diffs are also not displayed in the Pending Config window. But, the Side-by-side diff utility compares the diff in the two text files and does not leverage the internal logic used in the diff computation. As a result, the diff in default configurations are highlighted in red in the Side-by-side Comparison window.
Starting from Cisco DCNM Release 11.4(1), such diffs are not highlighted in the Side-by-side Comparison window. The auto-generated default configuration that is highlighted in the Running config window is not visible in the Expected config window.
Any configurations that are shown in the Pending Config window are highlighted in red in the Side-by-side Comparison window if the configurations are seen in the Running config window but not in the Expected config window. Also, any configurations that are shown in the Pending Config window are highlighted in green in the Side-by-side Comparison window if the configurations are seen in the Expected config window but not in the Running config window. If there are no configurations displayed in the Pending Config window, no configurations are shown in red in the Side-by-side Comparison window.
All freeform configurations have to strictly match the show running configuration output on the switch and any deviations from the configuration will show up as a diff during Save & Deploy. You need to adhere to the leading space indentations.
You can typically enter configuration snippets in DCNM using the following methods:
-
User-defined profile and templates
-
Switch, interface, overlay, and vPC freeform configurations
-
Network and VRF per switch freeform configurations
-
Fabric settings for Leaf, Spine, or iBGP configurations
Caution |
The configuration format should be identical to the show running configuration of the corresponding switch. Otherwise, any missing or incorrect leading spaces in the configuration can cause unexpected deployment errors and unpredictable pending configurations. If any unexpected diffs or deployment errors are displayed, check the user-provided or custom configuration snippets for incorrect values. |
If DCNM displays the "Out-of-Sync" status due to unexpected pending configurations, and this configuration is either unable to be deployed or stays consistent even after a deployment, perform the following steps to recover:
-
Check the lines of config highlighted under the Pending Config tab in the Config Preview window.
-
Check the same lines in the corresponding Side-by-side Comparison tab. This tab shows whether the diff exists in "intent", or "show run", or in both with different leading spaces. Leading spaces are highlighted in the Side-by-side Comparison tab.
-
If the pending configurations or switch with an out-of-sync status is due to any identifiable configuration with mismatched leading spaces in "intent" and "running configuration", this indicates that the intent has incorrect spacing and needs to be edited.
-
To edit incorrect spacing on any custom or user-defined policies, navigate to the switch and edit the corresponding policy:
-
If the source of the policy is UNDERLAY, you will need to edit this from the Fabric settings screen and save the updated configuration.
-
If the source is blank, it can be edited from the View/Edit policies window for that switch.
-
If the source of the policy is OVERLAY, but it is derived from a switch freeform configuration. In this case, navigate to the appropriate OVERLAY switch freeform configuration and update it.
-
If the source of the policy is OVERLAY or a custom template, perform the following steps:
-
Navigate to Administration > DCNM Server > Server Properties, set the template.in_use.check property to false. This allows the profiles or templates to be editable.
-
Edit the specific profile or template from the Control > Template Library edit window, and save the updated profile template with the right spacing.
-
Click Save & Deploy to recompute the diffs for the impacted switches.
-
After the configurations are updated, set the template.in_use.check property to true, as it slows down the performance of the DCNM system, specifically for Save & Deploy operations.
-
-
To confirm that the diffs have been resolved, click Save & Deploy after updating the policy to validate the changes.
Note |
DCNM checks only leading spaces, as it implies hierarchy of the command, especially in case of multi-command sequences. DCNM does not check any trailing spaces in command sequences. |
Example 1: Configuration Compliance in Switch Freeform Policy
Let us consider an example with an incorrect spacing in the Switch Freeform Config field.
The switch freeform policy is created as shown:
After deploying this policy successfully to the switch, DCNM persistently reports the following diffs:
After clicking the Side-by-side Comparison tab, you can see the cause of the diff. As seen below, the ip pim rp-address line has 2 leading spaces, while the running configuration has 0 leading spaces.
To resolve this diff, edit the corresponding Switch Freeform policy so that the spacing is correct.
After you save, you can use the Push Config or Save & Deploy option to re-compute diffs.
As shown below, the diffs are now resolved. The Side-by-side Comparison tab confirms that the leading spaces are updated.
Example 2: Resolving a Leading Space Error in Overlay Configurations
Let us consider an example with a leading space error that is displayed in the Pending Config tab.
In the Side-by-side Comparison tab, search for diffs line by line to understand context of the deployed configuration.
A matched count of 0 means that it is a special configuration that DCNM has evaluated to push it to the switch.
You can see that the leading spaces are mismatched between running and expected configurations.
Navigate to the respective freeform configs and correct the leading spaces, and save the updated configuration.
Navigate to the Fabric Builder window for the fabric and click Save & Deploy.
In the Config Deployment window, you can see that all the devices are in-sync.
Configuration Compliance in External Fabrics
With external fabrics, any Nexus switch can be imported into the fabric, and there is no restriction on the type of deployment. It can be LAN Classic, VXLAN, FabricPath, vPC, HSRP, etc. When switches are imported into an external fabric, the configuration on the switches is retained so that it is non-disruptive. Only basic policies such as the switch username and mgmt0 interface are created after a switch import.
In the external fabric, for any intent that is defined in the DCNM, configuration compliance (CC) ensures that this intent is present on the corresponding switch. If this intent is not present on the switch, CC reports an Out-of-Sync status. Additionally, there will be a Pending Config generated to push this intent to the switch to change the status to In-Sync. Any additional configuration that is on the switch but not in intent defined in DCNM, will be ignored by CC, as long as there is no conflict with anything in the intent.
When there is user-defined intent added on DCNM and the switch has additional configuration under the same top-level command, as mentioned earlier, CC will only ensure that the intent defined in DCNM is present on the switch. When this user defined intent on DCNM is deleted as a whole with the intention of removing it from the switch and the corresponding configuration exists on the switch, CC will report an Out-of-Sync status for the switch and will generate Pending Config to remove the config from the switch. This Pending Config includes the removal of the top-level command. This action leads to removal of the other out-of-band configurations made on the switch under this top-level command as well. If you choose to override this behavior, the recommendation is that, you create a freeform policy and add the relevant top-level command to the freeform policy.
Let us see this behavior with an example.
-
A switch_freeform policy defined by the user in DCNM and deployed to the switch.
-
Additional configuration exists under router bgp in Running config that does not exist in user-defined DCNM intent Expected config. Note that there is no Pending Config to remove the additional config that exists on the switch without a user defined intent on DCNM.
-
The Pending Config and the Side-by-side Comparison when the intent that was pushed earlier via DCNM is deleted from DCNM by deleting the switch_freeform policy that was created in the Step 1.
-
A switch_freeform policy with the top-level router bgp command needs to be created. This enables CC to generate the configuration needed to remove only the desired sub-config which was pushed from DCNM earlier.
-
The removed configuration is only the subset of the configuration that was pushed earlier from DCNM.
For interfaces on the switch in the external fabric, DCNM either manages the entire interface or does not manage it at all. CC checks interfaces in the following ways:
-
For any interface, if there is a policy defined and associated with it, then this interface is considered as managed. All configurations associated with this interface must be defined in the associated interface policy. This is applicable for both logical and physical interfaces. Otherwise, CC removes any out-of-band updates made to the interface to change the status to In-Sync.
-
Interfaces created out-of-band (applies for logical interfaces such as port-channels, sub interfaces, SVIs, loopbacks, etc.), will be discovered by DCNM as part of the regular discovery process. However, since there is no intent for these interfaces, CC will not report an Out-of-Sync status for these interfaces.
-
For any interface, there can always be a monitor policy associated with it in DCNM. In this case, CC will ignore the interface’s configuration when it reports the In-Sync or Out-of-Sync config compliance status.
-
Special Configuration CLIs Ignored for Configuration Compliance
The following configuration CLIs are ignored during configuration compliance checks:
-
Any CLI having 'username’ along with ‘password’
-
Any CLI that starts with ‘snmp-server user’
Any CLIs that match the above will not show up in pending diffs and clicking Save & Deploy in the Fabric Builder window will not push such configurations to the switch. These CLIs will not show up in the Side-by-side Comparison window also.
To deploy such configuration CLIs, perform the following procedure:
-
Select Control>Fabric Builder, click Tabular View, and select a switch in the Name column or select Control>Fabric Builder and right-click on the device.
-
Click View/Edit Policies and click on + to add a new policy. The Add Policy window comes up.
-
Add a PTI with the required configuration CLIs using the switch_freeform template and click Save.
-
Select the created policy and click Push Config to deploy the configuration to the switch(es).
Resolving Diffs for Case Insensitive Commands
By default, all diffs generated in DCNM while comparing intent, also known as Expected Configuration, and Running Configuration, are case sensitive. However, the switch has many commands that are case insensitive, and therefore it may not be appropriate to flag these commands as differences. These outlier cases are captured in the compliance_case_insensitive_clis.txt text file.
There could be additional commands not included in the existing compliance_case_insensitive_clis.txt file that should be treated as case insensitive. If the pending configuration is due to the differences of cases between the Expected Configuration in DCNM and the Running Configuration, you can configure DCNM to ignore these case differences as follows:
-
Modify the following file on the DCNM file system:
/usr/local/cisco/dcm/dcnm/model-config/compliance_case_insensitive_clis.txt
The sample entries in compliance_case_insensitive_clis.txt file are displayed as:
If newer patterns are detected during deployment, and they are triggering pending configurations, you can add these patterns to this file. The patterns need to be valid regex patterns.
This enables DCNM to treat the documented configuration patterns as case insensitive while performing comparisons.
-
Click Save & Deploy for fabrics to see the updated comparison outputs.
Resolving Config Compliance After Importing Switches
After importing switches in Cisco DCNM, configuration compliance for a switch can fail because of an extra space in the management interface (mgmt0) description field.
For example, before importing the switch:
interface mgmt0
description SRC=SDS-LB-LF111-mgmt0, DST=SDS-LB-SW001-Fa0/5
After importing the switch and creating a config profile:
interface mgmt0
description SRC=SDS-LB-LF111-mgmt0,DST=SDS-LB-SW001-Fa0/5
In this example, the space after the comma (,) is removed.
Navigate to Interface Manager and click the Edit icon after selecting the mgmt0 interface. Remove the extra space in the description.
Strict Configuration Compliance
From Cisco DCNM Release 11.3(1), strict configuration compliance checks for diff between the switch configuration and the associated intent and generates no commands for the configurations that are present on the switch but are not present in the associated intent. When you click Save and Deploy, switch configurations that are not present on the associated intent are removed. You can enable this feature by selecting the Enable Strict Config Compliance checkbox under the Advanced tab in the Add Fabric or Edit Fabric window. By default, this feature is disabled.
The strict config compliance feature is supported on the Easy Fabric templates - Easy_Fabric_11_1 and Easy_Fabric_eBGP. To avoid generating diff for commands that are auto-generated by the switch, such as vdc, rmon, and so on, a file that has a list of default commands is used by CC to ensure that diffs are not generated for these commands. This file is located at /usr/local/cisco/dcm/dcnm/model-config/strict_cc_exclude_clis.txt.
Note |
|
Example: Strict Configuration Compliance
Let us consider an example in which the feature telnet command is configured on a switch but is not present in the intent. In such a scenario, the status of the switch is displayed as Out-of-sync after a CC check is done.
Now, click Preview Config of the out-of-sync switch. As the strict config compliance feature is enabled, the no form of the feature telnet command appears under Pending Config in the Preview Config window.
Click the Side-by-side Comparison tab to display the differences between the running configuration and the expected configuration. Starting from Cisco DCNM Release 11.3(1), the Re-sync button is also displayed at the top right corner under the Side-by-side Comparison tab in the Preview Config window. Use this option to resynchronize DCNM state when there is a large scale out-of-band change, or if configuration changes do not register in the DCNM properly.
The re-sync operation does a full CC run for the switch and recollects “show run” and “show run all” commands from the switch. When you initiate the re-sync process, a progress message is displayed on the screen. During the re-sync, the running configuration is taken from the switch. The Out-of- Sync/In-Sync status for the switch is recalculated based on the intent defined in DCNM.
Now, close the Preview Config window and click Save and Deploy. The Strict configuration compliance feature then ensures that the running config on the switch does not deviate from the intent by pushing the no form of the feature telnet command to the switch. The diff between the configurations is highlighted. The diff other than the feature telnet command are default switch and boot configurations and are ignored by the strict CC check.
In Cisco DCNM Release 11.2(1) and earlier releases, you had to right-click on a switch in the Fabric builder window and select Deploy Config to display the Config Deployment window. You then had to click Preview Config for a specific switch to bring up the Preview Config window that displays the pending configuration for that switch. This leads to a scenario in which the user may think that the preview config is inadvertently being deployed on the switch. Starting from Cisco DCNM Release 11.3(1), you can right-click on a switch in the Fabric Builder window and select Preview Config to display the Preview Config window. This window displays the pending configuration that has to be pushed to the switch to achieve configuration compliance with the intent.
Custom freeform configurations can be added in DCNM to make the intended configuration on DCNM and the switch configurations identical. The switches are then in In-Sync status. For more information on how to add custom freeform configurations on DCNM, refer Enabling Freeform Configurations on Fabric Switches.
Enabling Freeform Configurations on Fabric Switches
In DCNM, you can add custom configurations through freeform policies in the following ways:
-
Fabric-wide
-
On all leaf, border leaf, and border gateway leaf switches in the fabric, at once.
-
On all spine, super spine, border spine, border super spine, border gateway spine and border switches, at once.
-
-
On a specific switch at the global level.
-
On a specific switch on a per Network or per VRF level.
Leaf switches are identified by the roles Leaf, Border, and Border Gateway. The spine switches are identified by the roles Spine, Border Spine, Border Gateway Spine, Super Spine, Border Super Spine, and Border Gateway Super Spine.
Note |
You can deploy freeform CLIs when you create a fabric or when a fabric is already created. The following examples are for an existing fabric. However, you can use this as a reference for a new fabric. |
Deploying Fabric-Wide Freeform CLIs on Leaf and Spine Switches
-
Click Control > Fabric Builder. The Fabric Builder screen comes up. A rectangular box represents each fabric.
-
Click the Edit Fabric icon (located on the top right part of the rectangular box) for adding custom configurations to an existing fabric. The Edit Fabric screen comes up.
(If you are creating a fabric for the first time, click Create Fabric).
-
Click the Advanced tab and update the following fields:
Leaf Freeform Config – In this field, add configurations for all leaf, border leaf, and border gateway leaf switches in the fabric.
Spine Freeform Config - In this field, add configurations for all Spine, Border Spine, Border Gateway Spine, Super Spine, Border Super Spine, and Border Gateway Super Spine switches in the fabric.
Note
Copy-paste the intended configuration with correct indentation, as seen in the running configuration on the Nexus switches. For more information, see Resolving Freeform Config Errors in Switches.
-
Click Save. The fabric topology screen comes up.
-
Click Save & Deploy at the top right part of the screen to save and deploy configurations.
Configuration Compliance functionality will ensure that the intended configuration as expressed by those CLIs are present on the switches and if they are removed or there is a mismatch, then it will flag it as a mismatch and indicate that the device is Out-of-Sync.
Incomplete Configuration Compliance - On some Cisco Nexus 9000 Series switches, in spite of configuring pending switch configurations using the Save & Deploy option, there could be a mismatch between the intended and switch configuration. To resolve the issue, add a switch_freeform policy to the affected switch (as explained in the Deploy Freeform CLIs on a Specific Switch section). For example, consider the following persistent pending configurations:
line vty
logout-warning 0
After adding the above configurations in a policy and saving the updates, click Save and Deploy in the topology screen to complete the deployment process.
To bring the switch back in-sync, you can add the above configuration in a switch_freeform policy saved and deployed onto the switch.
Deploying Freeform CLIs on a Specific Switch
-
Click Control > Fabric Builder. The Fabric Builder screen comes up.
-
Click on the rectangular box that represents the fabric. The Fabric Topology screen comes up.
Note
To provision freeform CLIs on a new fabric, you have to create a fabric, import switches into it, and then deploy freeform CLIs.
-
Right-click the switch icon and select the View/edit policies option.
The View/Edit Policies screen comes up.
-
Click +. The Add Policy screen comes up.
In the Priority field, the priority is set to 500 by default. You can choose a higher priority (by specifying a lower number) for CLIs that need to appear higher up during deployment. For example, a command to enable a feature should appear earlier in the list of commands.
-
From the Policy field, select switch_freeform.
-
Add or update the CLIs in the Freeform Config CLI box.
Copy-paste the intended configuration with correct indentation, as seen in the running configuration on the Nexus switches. For more information, see Resolving Freeform Config Errors in Switches.
-
Click Save.
After the policy is saved, it gets added to the intended configurations for that switch.
-
Close the policy screens. The Fabric Topology screen comes up again.
-
Right click the switch and click Deploy Config.
The Save & Deploy option can also be used for deployment. However, the Save & Deploy option will identify mismatch between the intended and running configuration across all fabric switches.
Pointers for switch_freeform Policy Configuration:
-
You can create multiple instances of the policy.
-
For a vPC switch pair, create consistent switch_freeform policies on both the vPC switches.
-
When you edit a switch_freeform policy and deploy it onto the switch, you can see the changes being made (in the Side-by-side tab of the Preview option).
Freeform CLI Configuration Examples
Console line configuration
This example involves deploying some fabric-wide freeform configurations (for all leaf, and spine switches), and individual switch configurations.
Fabric-wide session timeout configuration:
line console
exec-timeout 1
Console speed configuration on a specific switch:
line console
speed 115200
IP Prefix List/Route-map configuration
IP prefix list and route-map configurations are typically configured on border devices. These configurations are global because they can be defined once on a switch and then applied to multiple VRFs as needed. The intent for this configuration can be captured and saved in a switch_freeform policy. As mentioned earlier, note that the config saved in the policy should match the show run output. This is especially relevant for prefix lists where the NX-OS switch may generate sequence numbers automatically when configured on the CLI. An example snippet is shown below:
ACL configuration
ACL configurations are typically configured on specific switches and not fabric-wide (leaf/spine switches). When you configure ACLs as freeform CLIs on a switch, you should include sequence numbers. Else, there will be a mismatch between the intended and running configuration. A configuration sample with sequence numbers:
ip access-list ACL_VTY
10 deny tcp 172.29.171.67/32 172.29.171.36/32
20 permit ip any any
ip access-list vlan65-acl
10 permit ip 69.1.1.201/32 65.1.1.11/32
20 deny ip any any
interface Vlan65
ip access-group vlan65-acl in
line vty
access-class ACL_VTY in
If you have configured ACLs without sequence numbers in a switch_freeform policy, update the policy with sequence numbers as shown in the running configuration of the switch.
After the policy is updated and saved, right click the device and select the per switch Deploy Config option to deploy the configuration. Alternatively, use the Save and Deploy option in the fabric topology screen (within Fabric Builder) so that the fabric triggers Configuration Compliance and resolves the configuration mismatch.
Resolving Freeform Config Errors in Switches
Copy-paste the running-config to the freeform config with correct indentation, as seen in the running configuration on the NX-OS switches. The freeform config must match the running config. Otherwise, configuration compliance in DCNM marks switches as out-of-sync.
Let us see an example of the freeform config of a switch.
feature bash-shell
feature telemetry
clock timezone CET 1 0
# Daylight saving time is observed in Metropolitan France from the last Sunday in March (02:00 CET) to the last Sunday in October (03:00 CEST)
clock summer-time CEST 5 Sunday March 02:00 5 Sunday October 03:00 60
clock protocol ntp
telemetry
destination-profile
use-vrf management
The highlighted line about the daylight saving time is a comment that is not displayed in the show running config command output. Therefore, configuration compliance marks the switch as out-of-sync because the intent does not match the running configuration.
Let us check the running config in the switch for the clock protocol.
spine1# show run all | grep "clock protocol"
clock protocol ntp vdc 1
You can see that vdc 1 is missing from the freeform config.
In this example, let us copy-paste the running config to the freeform config.
Here is the updated freeform config:
feature bash-shell
feature telemetry
clock timezone CET 1 0
clock summer-time CEST 5 Sunday March 02:00 5 Sunday October 03:00 60
clock protocol ntp vdc 1
telemetry
destination-profile
use-vrf management
After you copy-paste the running config and deploy, the switch will be in-sync. When you click Save & Deploy, the Side-by-side Comparison tab in the Config Preview window provides you information about the difference between the defined intent and the running config.
Deploying Freeform CLIs on a Specific Switch on a Per VRF/Network basis
-
Click Control > VRFs. After choosing the appropriate fabric scope, the listing of the currently defined VRFs for the fabric shows up.
-
Create a new VRF by clicking the + button or select an existing VRF and click the Continue button on the top right.
-
The topology view for the fabric shows up. Switches to which the VRF is already deployed are highlighted in green. Other switches will be in gray color.
-
Select an individual switch. The VRF attachment form shows up listing the switch that is selected. In case of a vPC pair, both switches belonging to the pair will show up.
-
Under the CLI Freeform column, select the button labelled Freeform config. This option allows a user to specify additional configuration that should be deployed to the switch along with the VRF profile configuration.
-
Add or update the CLIs in the Free Form Config CLI box. Copy-paste the intended configuration with correct indentation, as seen in the running configuration on the Nexus switches. For more information, see Resolving Freeform Config Errors in Switches.
-
Click Save Config.
Note
The Freeform config button will be gray when there is no per VRF per switch config specified. The button will be blue when some config has been saved by the user.
After the policy is saved, Click Save on the VRF Attachment pop-up to save the intent to deploy the VRF to that switch. Ensure that the checkbox on the left next to the switch is checked.
-
Now, optionally, click Preview to look at the configuration that will be pushed to the switch.
-
Click Deploy to push the configuration to the switch.
The same procedure can be used to define a per Network per Switch configuration.
VMM Workload Automation
VMM workload automation is about the automation of network configuration in Cisco’s Nexus switches for workloads spawned in a VMware environment. Note that this is a preview feature in the Cisco DCNM Release 11.4(1).
You can also watch the video that demonstrates this automation. See Video: VMM Workload Automation in Cisco DCNM.
Overview of Network Objects in vCenter
VMM workload automation involves the mapping of network objects in vCenter to the network objects in DCNM. The following network objects in vCenter are considered:
-
Virtual Switch (VS): A regular VS runs in an ESXi host that performs software-based switching. A VS can have multiple port-groups (PG), where each PG has the network port configuration properties connecting to the network, like a VLAN. Each VS can have multiple uplink ports connecting to the leaf switches. Workloads spawned in the ESXi host can attach to the PG created in this VS.
-
Distributed Virtual Switch (DVS): A DVS is a virtual switch that spans across multiple ESXi hosts. Similar to a regular VS, a DVS has multiple port-groups referred to as Distributed Port Groups (DPG). A DPG has the network port configuration properties connecting to the network, like a VLAN. Each DVS can have multiple uplink ports, which can be connected to the leaf switches. Workloads spawned in any of the hosts that are members of the DVS, can attach to the DPG. Through this document, and in the config file, a DPG is also referred as Distributed Virtual Port Group (DV-PG).
Let us consider the following topology in DCNM:
-
There are four hosts in the setup with IP addresses:
-
172.28.11.33
-
172.28.10.154
-
172.28.10.151
-
172.28.10.155
-
-
A DVS named DSwitchNew spawns across hosts 172.28.10.154 and 172.289.10.155. This DVS has a DPG named DPGNew that isn’t shown in the figure. This DVS connects to switches n9k-leaf1 and n9k-leaf4 through uplink ports <vmnic3, vmnic1> and switch interfaces <e1/25, e1/7> respectively (not shown in the figure). A VLAN value of 571 is associated with the DPG DPGNew.
-
There’s also a regular vSwitch named vSwitch2 in host 172.28.11.33. This VS has a PG named DCNM_Inband. This VS connects to the leaf switch n9k-leaf5 through uplink port vmnic4 and switch interface e1/23. A VLAN value of 540 is associated with the PG DCNM_Inband.
How VMM Workload Automation Works
When workloads are spawned, they require provisioning in the network or fabric. The workloads spawned in the vCenter are either associated with a DPG or a PG. This DPG or PG in the VMWare should be mapped to a corresponding DCNM network. As an example in the earlier topology, when workloads are spawned in the host 172.28.11.33 with PG Inband, then network provisioning that includes overlay and underlay configuration needs to happen in the leaf switch n9k-leaf5 with the relevant interface e1/23.
For the network provisioning to happen, each network object in a vCenter (a DPG or a PG) should be mapped to a network object in DCNM. A network object in DCNM has the following characteristics:
-
VRF Name
-
VLAN ID
-
IPv4/IPv6 subnet and gateway information
-
Secondary IPv4/v6 and gateway information
-
BGP-EVPN configuration
A static mapping should be defined in a config file that maps the network object in a vCenter to a network object in DCNM. For information, see Configuration Files for VMM Workload Automation.
After the config file is populated, the workload automation module can be started. The module scans all the vCenters specified in the config file (conf.yml) and collects the following information for each vCenter:
-
List of DVS and the DPG configured in all data centers.
-
List of PGs configured in every host in all data centers.
-
For every DPG or PG specified in the conf file, it finds the configured VLANs and the directly connected neighbor switches along with its interface information.
-
For every <DVS, DPG> or <Host, PG> specified in the config file, it gets the associated network mapping in DCNM.
The module merges all the information and calls the DCNM APIs to provision or amend the networks in all the switches that are discovered as neighbors in one of the earlier steps.
Provisioning of a network or fabric uses DCNM top-down provisioning and it consists of the following steps:
-
Attaching the network configuration to the relevant interfaces of one or more switches that are discovered as neighbors. This attachment is done by the workload automation module.
-
After the configuration is attached, you can review the exact CLIs that are pushed to the switches.
-
After review, you can deploy the configuration to the switches. This deployment can be either done by the script based on the configuration file setting (default is False) or you can do it through DCNM. After this step, the configuration appears in the switches.
For more information, see https://pypi.org/project/vmm-workload-auto/.
Configuration Files for VMM Workload Automation
The following configuration files are used for VMM workload automation:
-
Global YML File (conf.yml): This file has global configuration and access or credential information of the DCNM and the vCenters. Also, the location of the CSV file for each DCNM is specified in this file. For more information, see Configuration Files for VMM Workload Automation.
-
CSV file (sample.csv): This file has the mapping of <DVS, DVS-PG> or <Host, PG> in vCenter to the Network name in DCNM. There’s a separate CSV file for each DCNM. For more information, see CSV File for Mapping Networks in vCenter and DCNM.
Configuration File for Mapping DCNM and vCenter
The configuration file (conf.yml) specifies the DCNM IP address, username, and password. For each DCNM, the list of vCenter information like the IP address, username, and password is also specified. Multiple DCNMs can be specified in this conf.yml file. For every DCNM instance, there’s an associated CSV file. The multi-DCNM case is only applicable when the script isn’t run in a DCNM, but run in a server that has connectivity to all the DCNMs and vCenters specified in the config file.
In the config file, the hierarchy of information that should be specified is as follows:
Global config parameters
DCNM1
DCNM1 config parameters including location of the CSV file
vCenter1
vCenter1 config parameters
…
vCenter2
vCenter2 config parameters
….
DCNM2
….
…
The location of this configuration file depends on the installation method of the VMM workload automation script. For more information, see Installing VMM Workload Automation Script. This file contains example entries. Modify it based on your environment.
The config file has the following entries:
LogFile: Specifies the name of the log file including the absolute path that will be used by the workload automation module for logging the errors and debug information. Make sure that the directory has write permission for creating the log file. For example, /tmp/workloadauto.log.
ListenPort: Specifies the port that the workload automation module uses to listen for the REST APIs, for example, 9590. Make sure that this port isn’t used by any other application. You can check the same by running the sudo netstat -tulpn command.
AutoDeploy: Specifies whether the script should automatically deploy the configuration in the switches after attaching the networks. By default, it’s set to False so that you can review the config and deploy it in the DCNM.
NwkMgr: Specifies the top-level section that contains the DCNM information. For multiple DCNM instances, repeat the fields with the appropriate values. For an example, see conf_multiple_dcnm.yml file that handles multiple DCNMs.
Ip: Specifies the IP address of DCNM, for example, 172.28.10.156.
User: Specifies the username used to log in to DCNM, for example, admin.
Password: Specifies the password of DCNM.
CsvFile: Specifies the absolute path of the location of the CSV file for this DCNM, for example, /etc/vmm_workload_auto/sample.csv.
ServerCntrlr: Specifies the information for the server controller, that is, vCenter/vSphere. For multiple vCenters that fall under this DCNM, this section repeats. For an example, refer the conf_multiple_vcenter.yml file, which contains multiple vCenters under a DCNM.
Ip: Specifies the IP address of the vCenter.
Type: Specifies the type of the server controller. The default is vCenter.
User: Refers to the username used to log in to the vCenter. For example, administrator@vsphere.local.
Password: Refers to the password for the vCenter.
The following example shows the contents of a conf.yml file:
LogFile: /tmp/workloadauto.log
ListenPort: 9590
AutoDeploy: false
NwkMgr:
- Ip: 172.28.10.151
User: admin
Password: C1sco_123
CsvFile: /etc/sample.csv
ServerCntrlr:
- Ip: 172.28.10.194
Type: vCenter
User: administrator@vsphere.local
Password: Cisc0!23
CSV File for Mapping Networks in vCenter and DCNM
The CSV file contains the mapping of the network object in vCenter to the network created in DCNM. This file has the following entries in CSV format, that is, comma-separated entries. The reason for having a CSV file is to specify the mapping between a PG (or DPG) of vSphere to the network name of DCNM. It’s a 1-1 mapping. However, since a PG or a DPG can’t be identified on its own (not unique), you need an extra DVS name or Hostname to map it.
The CSV file contains the following fields:
vCenter - Specifies the IP address of vCenter
Dvs - Specifies the name of the DVS.
Dvs_pg - Specifies the DVS PG (DPG) in the DVS
Host - Specifies the Host/Server (IP address)
Host_pg - Specifies the port-group in the host.
Fabric - Specifies the fabric in DCNM.
Network - Specifies the name of the network already created in DCNM.
The network object is identified by a unique pair of either <DVS, DVS_PG> or <Host, Host_PG>.
Consider the following example:
vCenter Params |
DCNM Params |
|||||
vCenter | DVS | DVPortGroup/ Network | ESXi Host | Port Group/ Network | Fabric Name | Network Name |
172.28.12.123 | DVS1 | DPG1 | Fab1 | Network 10 | ||
172.28.12.123 | DVS1 | DPG1 | Fab2 | Network 30 | ||
172.28.12.123 | 172.28.12.11 | PG10 | Fab1 | Network 20 | ||
172.28.12.123 | 172.28.12.12 | PG20 | Fab1 | Network 20 |
This table has the mapping for vCenter 172.28.12.123. It has four entries:
-
The first entry specifies that for DPG1 in DVS1, the network in DCNM is ‘Network10’ in fabric ‘Fab1’. There can be cases where in the hosts of the DVS can connect to switches in multiple fabrics. The network name in each fabric can be different, so you need the fabric name as well. The example in the table shows one such case in the second entry.
-
The second entry specifies the same <DVS1, DPG1> pair being mapped to Network 30 in the fabric ‘Fab2’.
-
The third entry specifies that for PG10 in the host 172.28.12.11, the network in DCNM is ‘Network20’ in fabric ‘Fab1’.
-
The fourth entry specifies that for PG20 in host 172.28.12.11, the network in DCNM is ‘Network20’ in fabric ‘Fab1’.
As seen in the earlier table, the network object is identified by a unique pair of either <DVS, DVS_PG> or <Host, Host PG>. If there’s a value specified for DVS, DVS_PG, then the values for <Host, Host_PG> are blank. In other words, <DVS, DVS_PG> and <Host, Host_PG) are mutually exclusive.
When the earlier table is specified in a CSV format, it appears as below in the CSV file:
172.28.12.123,DVS1,DPG1,,,Fab1,Network10
172.28.12.123,DVS1,DPG1,,,Fab2,Network30
172.28.12.123,,,172.28.12.11,PG10,Fab1,Network20
172.28.12.123,,,172.28.12.12,PG20,Fab1,Network20
Let’s consider more examples:
-
172.28.10.184,DSwitchPad,DSPad-PG2,,,DEF,MyNetwork_30000
This line in the CSV file specifies the IP address of vCenter as 172.28.10.184 and the <DVS, DVS PG> values are DSwitchPad, DSPad-PG2 respectively. Since the values for DVS, DVS-PG is specified, the values for Host, Host-PG are blank as seen in this example. The Fabric name is DEF and the network in DCNM is MyNetwork_30000.
-
172.28.10.184,,,172.28.11.33,Pad_Workload_Auto_Nwk,DEF,MyNetwork_60000
In this example, the values for <DVS, DVS-PG> is left blank and the values for <Host, Host_PG> is specified as 172.28.11.33 and Pad_Workload_Auto_Nwk respectively. The fabric in DCNM is DEF and the network name in DCNM is MyNetwork_60000.
An example CSV file is as follows:
vCenter,Dvs,Dvs_pg,Host,Host_pg,Fabric,Network
172.28.10.184,DSwitchNew,DPGNew,,,DEF,MyNetwork_30000
172.28.10.184,DSwitchNew,DPGNew,,,ABC,MyNetwork_30000
172.28.10.184,,,172.28.11.33,Pad_Workload_Auto_Nwk,DEF,MyNetwork_60000
Installing and Starting the VMM Workload Automation Module
You can install the VMM workload automation module by using the PIP install or the install script.
Using PIP Install
Before you begin
This installation method is for users who are familiar with pip install and know how to set up the proxy or handle cases when there’s a conflict in the python packages.
Procedure
Step 1 |
Decide whether you want to run this module in a virtual environment or on a physical server. If you decide to run this on a server, ensure that the you have the write permission for doing pip install. |
Step 2 |
Setup the http_proxy, https_proxy, and no_proxy appropriately. For example:
In this example, 172.28.10.0 specified in the no_proxy is the management subnet of DCNM. |
Step 3 |
Download and install the module from https://pypi.org/.
Similarly, you can uninstall the module using the command: pip3 uninstall vmm-workload-auto. |
Step 4 |
By default, the installation will happen in the following directories unless the you override by giving options in the pip command. The package is installed under: /usr/local/lib/python3.7/site-packages/vmm_workload_auto-0.1.1.dist-info The config files are installed under: /usr/local/lib/python3.7/site-packages/etc/vmm_workload_auto The source code is placed under: /usr/local/lib/python3.7/site-packages/workload_auto |
Step 5 |
Edit the config files in: /usr/local/lib/python3.7/site-packages/etc/vmm_workload_auto. For more information, see Configuration File for Mapping DCNM and vCenter. Make sure that the path of the CSV file specified in the conf.yml file is correct. |
Step 6 |
Start the VMM Workload automation module. The entry point for the python module is: /usr/local/bin/vmm_workload_auto. You can either run it as:
Or
If /usr/local/bin/ is already in $PATH. Provide the config file as a command-line option.
|
Using the Install Script
Using the install script is an alternate method for users who don’t want to use pip install. The install script performs the installation and starts the python module.
Procedure
Step 1 |
Navigate to https://pypi.org/project/vmm-workload-auto/ and download the latest .tar.gz file. |
Step 2 |
Untar it. For example:
|
Step 3 |
Modify the config/conf.yml and config/sample.csv according to your environment. |
Step 4 |
Run the setup script as “source setup.sh”. |
Step 5 |
The install script initially prompts the user to edit the conf.yaml and .csv files. The script will then prompt the user for proxy and other details. After everything is complete, the script installs the python packages and starts the module automatically. |
Step 6 |
For information about installing the script, see the Installation section in the README file at https://pypi.org/project/vmm-workload-auto/. |
Post Installation
After running the workload automation module, navigate to the DCNM Networks window and check whether the network attachments are completed. Review the configuration and deploy it, if AutoDeploy is set to false in the configuration file (conf.yml).
Additional Functionalities Using REST APIs
The workload automation module also provides the following REST APIs:
Note |
The REST APIs execute in a separate window after the VMM Workload automation module is running. Make sure that the automation module is running before running the REST APIs. |
-
Refresh - When the CSV file is changed, a refresh operation needs to be performed. This operation rereads the file and applies any new configuration if needed. The refresh API is as follows:
curl -XPOST http://127.0.0.1:{port}/workload_auto/refresh
-
Resync - When there’s any change in the DVS-PG, PG, VLAN, or neighbor switches, then a resync operation is needed. If there are any changes found, the configuration is reapplied accordingly. The resync API is as follows:
curl -XPOST http://127.0.0.1:{port}/workload_auto/resync
-
Clean - In order to clean up the network provisioning that was previously done using the module, a cleanup operation is needed. The clean API is as follows:
curl -XPOST http://127.0.0.1:{port}/workload_auto/clean
Events in vCenter
In the DCNM Release 11.4(1), real-time event processing isn’t done using the module. The various relevant events and its significance to this module are as follows:
Refresh
The refresh API enables the module to read the CSV file again and apply the network configuration to one or more the relevant switches. The refresh operation needs to be performed for the following events:
-
Add PG: Create an entry in the CSV file that specifies the associated network in DCNM for this PG. After the entry is added, call the refresh REST API.
-
Add DPG: Create an entry in the CSV file that specifies the associated network in DCNM for this DPG. After the entry is added, call the refresh REST API.
Resync
The resync API enables the module to discover the network objects and its associated properties again. The result of this resync operation is applying the network configuration to the new or changed switches or interfaces. Perform the resync operation for the following events:
-
Add host to a DVS.
-
Modify VLAN in a DPG or PG.
-
Change in topology: When any of the following information is changed, issue the Resync REST API to rediscover the topology and applying the REST API.
-
Neighbor switch change: This can happen if the attached leaf switch is replaced with a new switch or rewired to a different switch.
-
Interface change: This can happen due to rewiring to a different interface in the switch.
-
Host pNIC change.
-
Add an extra connection: This can happen when:
-
A regular interface in the host is made a port-channel by connecting an extra interface from the host to the switch.
-
An extra interface in the host connecting to a different switch forming a vPC pair.
-
-
No Action Required
You need not perform any action for the following events:
-
Add stand-alone host.
-
Add vSwitch.
-
Add DVS.
-
Delete DVS.
Mapping Change
The different scenarios for a mapping change in a CSV are as follows:
-
If a new mapping is added, run the refresh API after adding the mapping in the CSV file.
-
If the mapping between the vCenter network to the DCNM network needs to change, then run the clean REST API, modify the mapping the CSV file, and run the refresh REST API.
-
If the existing mapping needs to be deleted, then run the clean REST API, delete the mapping in the CSV file and run the refresh API.
Other Events
The other events and the operation that aren’t part of a category are as follows:
-
Host removed from DVS: When a host is removed from the DVS, the network configuration in the associated leaf switch and connected interface needs to be removed. This needs to be done for all the DPGs of this DVS. Navigate to DCNM and unattach the appropriate networks.
-
DPG or PG Delete: For all the network mappings specified in the spec file that are associated with this DPG or PG, remove the network configuration in the relevant switches and interfaces. Navigate to DCNM and unattach the appropriate networks.
-
Port Down or Switch Down: If the port or switch is permanently going to be offline, the configuration needs to be removed out of band. If the switch isn’t reachable from the host, but it’s still managed by DCNM, navigate to DCNM and unattach the appropriate networks.