topic Re: BL25P second CPU failure in BladeSystem - General

BL25P second CPU failure

Tolik1 — Fri, 06 Aug 2010 05:09:25 GMT

Hi,
Can someone help with the following:
We have 4 BL25P blades, all working OK with single CPU configuration (AMD Opt.252/275/280).
After second CPU upgrade - blade is powering up for 1 sec, shuts down, front panel "Health LED" is red and "VRM 2 FAIL" LED on the system board glows red. After removing second CPU - all fine. All systems have latest firmware and same simptoms. All CPUs are OK and tested on different systems.
Thank you

Re: BL25P second CPU failure

Jan Soska — Fri, 06 Aug 2010 15:40:18 GMT

Hello,
did you add some memory with CPU or at least move some memory to second CPU from first CPU slots? Remember AMD requires memory attached to each CPU as it has internal memory controlers...

Jan

Re: BL25P second CPU failure

Tolik1 — Mon, 09 Aug 2010 22:56:15 GMT

Hi,
Tried with 1,2 -5,6 banks and with all slots populated - same result.
(mem confirmed OK)

Re: BL25P second CPU failure

Jan Soska — Tue, 10 Aug 2010 06:01:50 GMT

Hello,
is it G1 or G2 bl25p ?
Do you have latest bios installed?

Jan

Re: BL25P second CPU failure

Tolik1 — Tue, 10 Aug 2010 06:11:57 GMT

It is G1 with latest updates

Re: BL25P second CPU failure

Proliant Watch — Tue, 10 Aug 2010 06:24:30 GMT

Check if it follows the following :

1. All DIMMs must be PC2-5300 Registered DDR2 DIMMs (667 MHz.)
2. Both DIMM slots in a memory bank must be populated.
3. Both DIMMs in a memory bank must be identical.
4. Processor 1 memory bank A must always be populated.
5. DIMM banks are active only when the corresponding processor socket is populated.
6. If installing only one bank on a processor, populate the bank physically farthest from the processor.
7. If mixing DIMM sizes on the same processor, the largest capacity DIMMs must be populated in the bank physically farthest from the processor. The smallest capacity DIMMs must be populated in the bank physically nearest the processor
e settings to the new card and there wont be ay data loss

Also, try swapping the processors... good luck!!

Re: BL25P second CPU failure

Tolik1 — Tue, 10 Aug 2010 06:29:27 GMT

Everything was checked few times as per previous comment, did -not work...

Re: BL25P second CPU failure

Jan Soska — Tue, 10 Aug 2010 06:39:19 GMT

If you start server again with only 1 cpu, is there something in ILO or IML log?

Jan

Re: BL25P second CPU failure

Proliant Watch — Tue, 10 Aug 2010 06:41:19 GMT

Can you swap VRM?

Check this thread also....

http://forums11.itrc.hp.com/service/forums/questionanswer.do?admit=109447626+1281425927972+28353475&threadId=1382396

Re: BL25P second CPU failure

Proliant Watch — Tue, 10 Aug 2010 06:43:02 GMT

VRM... Voltage regulator module... now its known as PPM.... Try swapping or swap with another server....

Re: BL25P second CPU failure

Jan Soska — Tue, 10 Aug 2010 06:44:54 GMT

probably try this:
at http://h20000.www2.hp.com/bizsupport/TechSupport/DocumentIndex.jsp?contentType=SupportManual〈=en&cc=us&docIndexId=64180&taskId=101&prodTypeId=3709945&prodSeriesId=398225 is HP ProLiant Servers Troubleshooting Guide. At page 161 starts description about p Class blades and their error codes, maybe it helps...

Jan

Re: BL25P second CPU failure

Tolik1 — Wed, 11 Aug 2010 03:09:30 GMT

VRMs are soldered on board (not removable)

Re: BL25P second CPU failure

Mike Texter — Tue, 22 Feb 2011 20:37:51 GMT

Confirmed today that the system board is a lost cause if one or both of the VRM(#)_FAIL lights is on. All other components may still test out fine in other systems, but those VRM lights are a sign of instantaneous death.