Contents
Using basic file transfer services, you can configure a router as a Trivial File Transfer Protocol (TFTP) or Reverse Address Resolution Protocol (RARP) server, configure the router to forward extended BOOTP requests over asynchronous interfaces, and configure rcp, rsh, and FTP.
Your software release may not support all the features documented in this module. For the latest caveats and feature information, see Bug Search Tool and the release notes for your platform and software release. To find information about the features documented in this module, and to see a list of the releases in which each feature is supported, see the feature information table.
Use Cisco Feature Navigator to find information about platform support and Cisco software image support. To access Cisco Feature Navigator, go to www.cisco.com/go/cfn. An account on Cisco.com is not required.
You should have at least a basic familiarity with the Cisco IOS environment and the command-line interface.
You should have at least a minimal configuration running on your system.
You must have your network up and running, with Cisco IOS Release 12.2 or a later release installed.
Some of the Cisco IOS configuration commands are only available on certain router platforms, and the command syntax may vary on different platforms.
It is too costly and inefficient to have a machine that acts only as server on every network segment. However, when you do not have a server on every segment, your network operations can incur substantial time delays across network segments. You can configure a router to serve as a RARP or TFTP server to reduce costs and time delays in your network while allowing you to use your router for its regular functions.
Typically, a router that is configured as a TFTP or RARP server provides other routers with system image or router configuration files from its Flash memory. You can also configure the router to respond to other types of service requests, such as requests.
As a TFTP server host, the router responds to TFTP Read Request messages by sending a copy of the system image contained in ROM or one of the system images contained in Flash memory to the requesting host. The TFTP Read Request message must use one of the filenames that are specified in the configuration.
Note | For the Cisco 7000 family, the filename used must represent a software image that is present in Flash memory. If no image resides in Flash memory, the client router will boot the server’s ROM image as a default. |
Flash memory can be used as a TFTP file server for other routers on the network. This feature allows you to boot a remote router with an image that resides in the Flash server memory.
Some Cisco devices allow you to specify one of the different Flash memory locations (bootflash:, slot0:, slot1:, slavebootflash:, slaveslot0:, or slaveslot1:) as the TFTP server.
Reverse Address Resolution Protocol (RARP) is a protocol in the TCP/IP stack that provides a method for finding IP addresses based on MAC (physical) addresses. This functionality is the reverse of broadcasting Address Resolution Protocols (ARPs), through which a host can dynamically discover the MAC-layer address corresponding to a particular IP network-layer address. RARP makes diskless booting of various systems possible (for example, diskless workstations that do not know their IP addresses when they boot, such as Sun workstations or PCs on networks where the client and server are on separate subnets). RARP relies on the presence of a RARP server with cached table entries of MAC-layer-to-IP address mappings.
You can configure a Cisco router as a RARP server. This feature enables the Cisco IOS software to answer RARP requests.
Remote shell (rsh) gives users the ability to execute commands remotely. Remote copy (rcp) allows users to copy files to and from a file system residing on a remote host or server on the network. Cisco’s implementation of rsh and rcp interoperates with the industry standard implementations. Cisco uses the abbreviation RCMD (Remote Command) to indicate both rsh and rcp.
You can specify the source interface for RCMD (rsh and rcp) communications. For example, the router can be configured so that RCMD connections use the loopback interface as the source address of all packets leaving the router. Specifying the source-interface is most commonly used to specify a loopback interface. This allows you to associate a permanent IP address with RCMD communications. Having a permanent IP address is useful for session identification (remote device can consistently idendify the origin of packets for the session). A “well-known” IP address can also be used for security purposes, as you can then create access lists on remote devices which include the address.
As a basic security check, the Cisco IOS software does a reverse lookup of the client IP address using DNS for the remote command (rcmd) applications (rsh and rcp). This check is performed using a host authentication process.
When enabled, the system records the address of the requesting client. That address is mapped to a host name using DNS. Then a DNS request is made for the IP address for that host name. The IP address received is then checked against the original requesting address. If the address does not match with any of the addresses received from DNS, the rcmd request will not be serviced.
This reverse lookup is intended to help protect against “spoofing.” However, please note that the process only confirms that the IP address is a valid routable address; it is still possible for a hacker to spoof the valid IP address of a known host.
You can use rsh (remote shell) to execute commands on remote systems to which you have access. When you issue the rsh command, a shell is started on the remote system. The shell allows you to execute commands on the remote system without having to log in to the target host.
You do not need to connect to the system, router, or access server and then disconnect after you execute a command if you use rsh. For example, you can use rsh to remotely look at the status of other devices without connecting to the target device, executing the command, and then disconnecting. This capability is useful for looking at statistics on many different routers. Configuration commands for enabling rsh use the acronym “rcmd”, which is short for “remote command”.
To gain access to a remote system running rsh, such as a UNIX host, an entry must exist in the system’s .rhosts file or its equivalent identifying you as a user who is authorized to execute commands remotely on the system. On UNIX systems, the .rhosts file identifies users who can remotely execute commands on the system.
You can enable rsh support on a router to allow users on remote systems to execute commands. However, our implementation of rsh does not support an .rhosts file. Instead, you must configure a local authentication database to control access to the router by users attempting to execute commands remotely using rsh. A local authentication database is similar to a UNIX .rhosts file. Each entry that you configure in the authentication database identifies the local user, the remote host, and the remote user.
The remote copy (rcp) commands rely on the rsh server (or daemon) on the remote system. To copy files using rcp, you do not need to create a server for file distribution, as you do with TFTP. You need only to have access to a server that supports the remote shell (rsh). (Most UNIX systems support rsh.) Because you are copying a file from one place to another, you must have read permission on the source file and write permission in the destination directory. If the destination file does not exist, rcp creates it for you.
Although Cisco’s rcp implementation emulates the functions of the UNIX rcp implementation--copying files among systems on the network--Cisco’s command syntax differs from the UNIX rcp command syntax. The Cisco IOS software offers a set of copy commands that use rcp as the transport mechanism. These rcp copy commands are similar in style to the Cisco IOS TFTP copy commands, but they offer an alternative that provides faster performance and reliable delivery of data. These improvements are possible because the rcp transport mechanism is built on and uses the Transmission Control Protocol/Internet Protocol (TCP/IP) stack, which is connection-oriented. You can use rcp commands to copy system images and configuration files from the router to a network server and vice versa.
You can also enable rcp support to allow users on remote systems to copy files to and from the router.
If you do not specify the/user keyword and argument, the Cisco IOS software sends a default remote username. As the default value of the remote username, the software sends the remote username associated with the current tty process, if that name is valid. If the tty remote username is invalid, the software uses the router host name as the both the remote and local usernames.
The rcp protocol requires a client to send a remote username on each rcp request to a server. When you copy a configuration file from a server to the router using rcp, the Cisco IOS software sends the first valid username in the following list:
The username set by the iprcmdremote-username command, if the command is configured.
The remote username associated with the current tty (terminal) process. For example, if the user is connected to the router through Telnet and was authenticated through the username command, the router software sends the Telnet username as the remote username.
Note | In Cisco products, ttys are commonly used in access servers. The concept of tty originated with UNIX. For UNIX systems, each physical device is represented in the file system. Terminals are called tty devices , which stands for teletype , the original UNIX terminal. |
The router host name.
For bootcommands using rcp, the software sends the router host name; you cannot explicitly configure the remote username.
For the rcp copy request to execute successfully, an account must be defined on the network server for the remote username.
If you are writing to the server, the rcp server must be properly configured to accept the rcp write request from the user on the router. For UNIX systems, you must add an entry to the .rhosts file for the remote user on the rcp server. For example, if the router contains the following configuration lines.
hostname Rtr1 ip rcmd remote-username User0
and the router’s IP address translates to Router1.company.com, then the .rhosts file for User0 on the rcp server should contain the following line:
Router1.company.com Rtr1
Refer to the documentation for your rcp server for more details.
If the server has a directory structure, the configuration file or image is written or copied relative to the directory associated with the remote username on the server. Use the iprcmdremote-username command to specify which directory on the server to use. For example, if the system image resides in the home directory of a user on the server, you can specify that user’s name as the remote username.
If you copy the configuration file to a personalcomputer used as a file server, the computer must support rsh.
The server and client router must be able to reach each other before the TFTP function can be implemented. Verify this connection by testing the connection between the server and client router (in either direction) using the pinga.b.c.d command (where a.b.c.d is the address of the client device). After the ping command is issued, connectivity is indicated by a series of exclamation points (!), while a series of periods (.) plus [timed out] or [failed] indicates that the connection attempt failed. If the connection fails, reconfigure the interface, check the physical connection between the Flash server and client router, andping again.
After you verify the connection, ensure that a TFTP-bootable image is present on the server. This is the system software image the client router will boot. Note the name of this software image so you can verify it after the first client boot.
Caution | For full functionality, the software image sent to the client must be the same type as the ROM software installed on the client router. For example, if the server has X.25 software, and the client does not have X.25 software in ROM, the client will not have X.25 capabilities after booting from the server’s image in Flash memory. |
In the following example, the system can use TFTP to send copies of the Flash memory file version-10.3 in response to a TFTP Read Request for that file. The requesting host is checked against access list 22.
tftp-server flash version-10.3 22
In the following example, the system can use TFTP to send a copy of the ROM image gs3-k.101in response to a TFTP Read Request for the gs3-k.101 file:
tftp-server rom alias gs3-k.101
In the following example, a router sends a copy of the file gs7-k.9.17 in Flash memory in response to a TFTP Read Request. The client router must reside on a network specified by access list 1. Thus, in the example, the any clients on network 172.16.101.0 are permitted access to the file.
Server# configure terminal Enter configuration commands, one per line. End with CTRL/Z Server(config)# tftp-server flash gs7-k.9.17 1 Server(config)# access-list 1 permit 172.16.101.0 0.0.0.255 Server(config)# end Server# copy running-config startup-config [ok] Server#
The TFTP session can sometimes fail. TFTP generates the following special characters to help you determine why a TFTP session fails:
An “E” character indicates that the TFTP server received an erroneous packet.
An “O” character indicates that the TFTP server received an out-of-sequence packet.
A period (.) indicates a timeout.
For diagnosing any undue delay in the transfer, the output is useful. For troubleshooting procedures, refer to the Internetwork Troubleshooting Guide publication.
To configure the client router to first load a system image from the server, and as a backup, to configure the client router to load its own ROM image if the load from a server fails, complete the tasks in this section:
In the following example, the router is configured to boot from a specified TFTP server:
Client# configure terminal Enter configuration commands, one per line. End with CTRL/Z Client(config)# no boot system Client(config)# boot system c5300-js-mz.121-5.T.bin 172.16.1.1 Client(config)# boot system rom Client(config)# config-register 0x010F Client(config)# end Client# copy running-config startup-config [ok] Client# reload
In this example, the nobootsystem command invalidates all otherbootsystem commands currently in the configuration memory, and any bootsystem commands entered after this command will be executed first. The second command, bootsystemfilename address , tells the client router to look for the file c5300-js-mz.121-5.T.bin on the TFTP server with an IP address of 172.16.111.111. Failing this, the client router will boot from its system ROM in response to the bootsystemrom command, which is included as a backup in case of a network problem. The copyrunning-configstartup-config command copies the configuration to the startup configuration, and thereloadcommand boots the system.
Note | The system software to be booted from the server must reside in Flash memory on the server. If it is not in Flash memory, the client router will boot the server’s system ROM. |
The following example shows sample output of the showversion command after the router has rebooted:
Device> show version Cisco Internetwork Operating System Software Cisco IOS (tm) 5300 Software (C5300-JS-M), Version 12.1(5)T, RELEASE SOFTWARE (fc1) Copyright (c) 1986-2000 by Cisco Systems, Inc. Compiled Sat 11-Nov-00 03:03 by joe Image text-base: 0x60008958, data-base: 0x611C6000 ROM: System Bootstrap, Version 11.2(9)XA, RELEASE SOFTWARE (fc2) BOOTFLASH: 5300 Software (C5300-BOOT-M), Version 12.0(7)T, RELEASE SOFTWARE (f) Router uptime is 8 weeks, 4 days, 22 hours, 36 minutes System returned to ROM by power-on System restarted at 00:37:38 UTC Thu Feb 22 2001 System image file is "flash:c5300-js-mz.121-5.T.bin" . . . Configuration register is 0x010F
The important information in this example is contained in the first line “Cisco IOS (tm)..” and in the line that begins “System image file....” The “Cisco IOS (tm)...” line shows the version of the operating system in NVRAM. The “System image file....” line show the filename of the system image loaded from the TFTP server.
After the system reloads, you should use the showversion EXEC mode command to verify that the system booted the desired image.
Caution | Using the nobootsystem command, as in the following example, will invalidate all other boot system commands currently in the client router system configuration. Before proceeding, determine whether the system configuration stored in the client router should first be saved (uploaded) to a TFTP file server so you have a backup copy. |
The figure below illustrates a network configuration in which a router is configured to act as a RARP server for a diskless workstation. In this example, the Sun workstation attempts to resolve its MAC (hardware) address to an IP address by sending a SLARP request, which is forwarded by the router to the Sun server.
Router A has the following configuration:
! Allow the router to forward broadcast portmapper requests ip forward-protocol udp 111 ! Provide the router with the IP address of the diskless sun arp 172.30.2.5 0800.2002.ff5b arpa interface Gigabitethernet 0/0 ! Configure the router to act as a RARP server, using the Sun Server's IP ! address in the RARP response packet. ip rarp-server 172.30.3.100 ! Portmapper broadcasts from this interface are sent to the Sun Server. ip helper-address 172.30.3.100
The Sun client and server’s IP addresses must use the same major network number because of a limitation with the current SunOS rpc.bootparamd daemon.
In the following example, an access server is configured to act as a RARP server.
! Allow the access server to forward broadcast portmapper requests ip forward-protocol udp 111 ! Provide the access server with the IP address of the diskless sun arp 172.30.2.5 0800.2002.ff5b arpa interface Gigabitethernet 0/0 ! Configure the access server to act as a RARP server, using the Sun Server's ! IP address in the RARP response packet. ip rarp-server 172.30.3.100 ! Portmapper broadcasts from this interface are sent to the Sun Server. ip helper-address 172.30.3.100
To configure the router so that RCMD connections use the loopback interface as the source address of all packets leaving the router, specify the interface associated with RCMD communications by completing the task in this section:
DNS Reverse Lookup for rcmd is enabled by default. You can disable the DNS check for RCMD (rsh and rcp) access by completing the task in this section:
To configure the router to allow remote user to execute commands using rsh, complete the tasks in this section:
The following example shows how to add two entries for remote users to the authentication database, and enable a router to support rsh commands from remote users:
ip rcmd remote-host Router1 172.16.101.101 rmtnetad1 ip rcmd remote-host Router1 172.16.101.101 netadmin4 enable ip rcmd rsh-enable
The users, named rmtnetad1 and netadmin4 , are both on the remote host at IP address 172.16.101.101. Although both users are on the same remote host, you must include a unique entry for each user. Both users are allowed to connect to the router and remotely execute rsh commands on it after the router is enabled for rsh. The user named netadmin4 is allowed to execute privileged EXEC mode commands on the router. Both authentication database entries give the router’s host name Router1 as the local username. The last command enables the router for to support rsh commands issued by remote users.
To execute a command remotely on a network server using rsh, use the following commands in user EXEC mode:
Command or Action | Purpose |
---|
The following example executes the “ls -a” command in the home directory of the user sharon on mysys.cisco.com using rsh:
Device# enable Device# rsh mysys.cisco.com /user sharon ls -a . .. .alias .cshrc .emacs .exrc .history .login .mailrc .newsrc .oldnewsrc .rhosts .twmrc .xsession jazz Device#
To configure the Cisco IOS software to support incoming rcp requests, use the following commands in global configuration mode:
Command or Action | Purpose | |||||
---|---|---|---|---|---|---|
Step 1 |
enable
Example: Device> enable |
Enables privileged EXEC mode. | ||||
Step 2 |
configure
terminal
Example: Device# configure terminal |
Enters global configuration mode. | ||||
Step 3 |
ip
rcmd
remote-host
local-username
{ip-address |
host }
remote-username
[enable[level ]]
Example: Device(config)# ip rcmd remote-host Router1 172.16.101.101 netadmin3 |
Create an entry in the local authentication database for each remote user who is allowed to execute rcp commands.
| ||||
Step 4 |
ip
rcmd
rcp-enable
Example: Device(config)# ip rcmd rcp-enable |
Enable the software to support incoming rcp requests. |
The following example shows how to add two entries for remote users to the authentication database and then enable the software to support remote copy requests from remote users. The users, named netadmin1 on the remote host at IP address 172.16.15.55 and netadmin3 on the remote host at IP address 172.16.101.101, are both allowed to connect to the router and remotely execute rcp commands on it after the router is enabled to support rcp. Both authentication database entries give the host name Router1 as the local username. The last command enables the router to support for rcp requests from remote users.
ip rcmd remote-host Router1 172.16.15.55 netadmin1 ip rcmd remote-host Router1 172.16.101.101 netadmin3 ip rcmd rcp-enable
To override the default remote username sent on rcp requests, use the following command in global configuration mode:
Command or Action | Purpose | |||
---|---|---|---|---|
Step 1 |
enable
Example: Device> enable |
Enables privileged EXEC mode. | ||
Step 2 |
configure
terminal
Example: Device# configure terminal |
Enters global configuration mode. | ||
Step 3 |
ip
rcmd
remote-username
username
Example: Device(config)# ip rcmd remote-username sharon |
Specifies the remote username.
|
To configure a router to transfer files between systems on the network using the File Transfer Protocol (FTP), complete the tasks in this section to configure the FTP characteristics:
Command or Action | Purpose | |
---|---|---|
Step 1 |
enable
Example: Device> enable |
Enables privileged EXEC mode. |
Step 2 |
configure
terminal
Example: Device# configure terminal |
Enters global configuration mode. |
Step 3 |
ip
ftp
username
string
Example: Device(config)# ip ftp username zorro |
Specifies the user name to be used for the FTP connection. |
Step 4 |
ip
ftp
password
[type ]
password
Example: Device(config)# ip ftp password sword |
Specifies the password to be used for the FTP connection. |
Step 5 | Do one of the
following:
Example: Device(config)# ip ftp passive |
Configures the router to only use passive-mode FTP connections. or Allows all types of FTP connections (default). |
Step 6 |
ip
ftp
source-interface
interface
Example: Device(config)# ip ftp source-interface to1 |
Specifies the source IP address for FTP connections. |
The following example demonstrates how to capture a core dump using the Cisco IOS FTP feature. The router accesses a server at IP address 192.168.10.3 with login name zorro and password sword. The default passive-mode FTP is used, and the server is accessed using Token Ring interface to1 on the router where the core dump will occur:
ip ftp username zorro ip ftp password sword ip ftp passive ip ftp source-interface to1 ! The following command allows the core-dump code to use FTP rather than TFTP or RCP exception protocol ftp ! The following command identifies the FTP server ! 192.168.10.3 crashes exception dump 192.168.10.3
Copyright © 2015, Cisco Systems, Inc. All rights reserved.