Field Notice: FN72072 - UCS S3260 M5 Server Power-On Failure - Replace on Failure

A limited number of Unified Computing System (UCS) S3260 M5 server nodes (UCS-S3260-M5SRB) might fail to power on due to a damaged capacitor.

A subset of S3260 M5 server nodes shipped before February 2020 with a specific Voltage Regulator Module (VRM) might fail over time due to mechanical stress. This can result in a failure to power on.

This issue is isolated to a specific VRM component which is no longer in use.

Systems with susceptible components can be identified by the server serial number. See the Serial Number Validation section for the validation link.

The Standalone Cisco Integrated Management Controller (IMC) System Event Log (SEL) will report:

• Platform alert POWER_ON_FAIL #0x34 | Predictive Failure asserted | Asserted
• Critical FRU_MB POWER_ON_FAIL: Platform sensor for FRU_MB, Predictive Failure asserted

UCS Manager integrated systems will report these faults:

• Motherboard of server [X/Y] (service profile: ) power: failed
• Unable to change server power state-MC Error(-20)

The Cisco IMC CLI output of the power status command will report VDD-Power-Good “inactive” and Power-On-Fail “active”, as shown in this example:

   Power Status:
   OP: [ status ]
   Power-State:                 [ on ]
   Master-State:                [ Master ]
   VDD-Power-Good:              [ inactive ]
   Power-On-Fail:               [ active ]  
   Power-Ctrl-Lock:             [ unlocked ]
   Power-System-Status:         [ Good ]
   Front-Panel Power Button:    [ Disabled ]
   Front-Panel Reset Button:    [ Disabled ]
   Source of Last Power Change: [ No Transition ]

Technical support logs can be checked in order to verify that the reason for the Power-On-Fail matches this specific issue. If the serial number is affected and you have experienced these symptoms, follow the normal Return Materials Authorization (RMA) process to request a replacement.

This is a hardware issue with no software workaround. Failed nodes should be replaced. The server serial numbers should be checked for exposure with the Serial Number Validation tool. See the Serial Number Validation section for the link.

If nodes are identified as requiring replacement, also validate if a Trusted Platform Module (TPM) is currently installed. See the How to Identify Affected Products section for more information.

Impacted servers can be identified by the serial number. Verification methods vary slightly based on management mode - UCS Manager or Standalone Cisco IMC software.

CLI - UCS Manager

Use SSH to connect to your UCS Manager CLI and enter this command in order to list all server inventory.

6454-FI-B# show server inventory
    2/1     UCS-S3260-M5SRB      V01          FCH22207VKZ          Equipped                   786432

GUI - UCS Manager

In UCS Manager, click the Equipment tab and navigate to the S3260 chassis. For each server within the chassis, click the General tab. Check the Properties section for the serial number details.

CLI - Standalone Cisco IMC

Use SSH to connect to your Cisco IMC CLI and enter this command:

    S3260# show server
    Server ID Serial Number PID
    --------- ------------- ----------------
    1         FCH221872CC   UCS-S3260-M5SRB

GUI – Standalone Cisco IMC

Open the menu from the top-left corner and navigate to the Compute section. Choose a server and click the General tab. In the Server Properties section, confirm the serial number.

TPM Validation

If nodes are identified as requiring replacement, also validate if a TPM is currently installed. If so, a new TPM of the matching PID must also be requested in your RMA.

CLI - UCS Manager

Use SSH to connect to your UCS Manager CLI and enter these commands in order to confirm the TPM status and PID for each impacted server. Confirm if a TPM is populated, Model (PID), and Serial Number.

scope server X/Y (X = chassis number, Y= blade number for each server in question)
scope tpm 1
show detail

Example:

`scope server 3/1`
`scope tpm 1;show detail`
Trusted Platform Module:
    Enabled Status: Enabled
    Active Status: Activated
    Ownership: Owned
    Tpm Revision: 2
    Model: UCSX-TPM2-002

GUI - UCS Manager

In UCS Manager, click the Equipment tab and navigate to the S3260 chassis. For each server within the chassis, click the Inventory tab, then the Motherboard tab. Scroll down and expand the TPM section, which will show if a TPM is installed. Confirm if a TPM is populated, Model (PID), and Serial Number.

CLI - Standalone Cisco IMC

Use SSH to connect to your Cisco IMC CLI and run these commands:

scope chassis
scope server X (1 and/or 2)
show tpm-inventory

Confirm if a TPM is populated, Model (PID), and Serial Number.

 
GUI – Standalone Cisco IMC

Open the menu from the top-left corner and navigate to the Compute section. Choose a server and click the Inventory tab, then the TPM tab. Confirm if a TPM is populated, Model (PID), and Serial Number.