CPU thermal trip errors in the BMC system event log - IBM System x3450

Resolving The Problem

Source

RETAIN tip: H196750

Symptom

The Baseboard Management Controller (BMC) System Event Log (SEL) contains repeated entries for Central Processing Unit (CPU) Thermal Trip events.

Affected configurations

The system may be any of the following IBM servers:

• System x3450, type 7948, any model

This tip is not software specific.
This tip is not option specific.

Additional information

There may be several reasons for this issue:

1. There is not enough thermal grease between the processor and the heatsinks, or the heatsinks are not completely screwed down on to the system board.
   
   
2. The processors were changed at some point from the time the replacement part numberSDR utility was run. It is recommended to re-run the replacement part numberSDR10 utility to reload the SDRs for these processors.
   
   
3. Make sure that the air duct is properly installed prior to powering on the system.
   
   
4. Make sure that the chassis cover is fully installed prior to powering on the system.
   
   
5. When a CPU thermal trip error occurs, it is logged to the SEL and recorded in a CPU history log that Basic Input/Output System (BIOS) maintains. As long as the event remains in the CPU history log, the error will persist with subsequent reboots. Even if the user swaps out the processors with new ones, as long as the original event is still present in the history log, the error will persist and the CPU will remain offline. The only way to clear the error is to clear the history log. This is done by setting the "Processor Retest" option in the F2 BIOS Setup utility.