Please analyze K class chassis codes

[Stefan Farrelly]

Hi,

could someone from HP please analyze these chassis codes from a K class crash overnight and let us know which piece of hardware is most likely at fault ?

HPMC Chassis Codes = 0xcbf0 0x5008 0x5408 0x550f 0xcbfb

Thanks,

Stefan

[Stefan Saliba]

0xcbf0 HPMC Occured

0x5008 Processor memory Bus Broad fault reported from CPU in slot 0 and Bus number 0 i.e runway bus)

0x5408 Processor memory Bus Broad fault reported from system board and Bus number 0 i.e runway bus)

0x550f TLB fault in the I/O adapter reported from system baord on bus 0

0xcbfb End of HPMC


check dump if you have and try running q4.

It seems some I/O adapter failed. Could be the one on the system board.

Check OLD_syslog.log and tombstones files as well.


Good luck stefan

[Chris Wilshaw]

Ignoring the 1st and last set of codes, I believe that 0x5 indicates a CPU or system board error. (0x2 would [I think] mean memory)

An HP engineer would be able to qualify this further from the other codes. In the past, I've tried to obtain a list of these codes, but there doesn't appear to be one available.

[Michael Tully]

Hi Stefan,

Have a look here:

http://forums.itrc.hp.com/cm/components/FileAttachment/0,,0x1f8e84534efbd5118ff40090279cd0f9,00.htm

Cheers
Michael

[Paula J Frazer-Campbell]

Hi Stefan

Have you got this yet ?

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x148f93e260b0d611abdb0090277a778c,00.html

Lots of good help

Paula

[Stefan Farrelly]

Hi all,

Ive had the chassis codes analyzed by someone at HP (Cheers Tony!) with the chassis code analzyer and the results are;

Decode of HPMC Hex Codes: (Click on codes with hyperlinks for more information)

cbf0: HPMC occurred - this code normally comes first.
5008: Broadcast Error : Processor 0
5408: Broadcast Error : IOA0
550f: U2 TLB error or invalid PDIR entry : IOA1
cbfb: PDCE_HPMC processing complete, branching to OS_HPMC

Now, the 550f is the important one, and according to further analysis this is most likely a SOFTWARE problem - not hardware. This has happened before and we did replace the SCSI card, but now its recurred im convinced its software, questions is what to do about it, we already have the Mar2002 bundles loaded. I shall investigate further.

Michael - that link was handy, thanks, but not detailed enough - it indicated a hardware problem!

Paula - those new recover handbooks look good, but like other, I cant access them - even with a support contract and system handles registered - something wrong with it if you ask me, hopefully they will fix it soon.

Cheers,

stefan

[Stefan Saliba]

With all due respect Mr. Farrely but I could have given you what HP gave you.

But do you understand it ???

I gave just the same thing but took the pains to decrypt and write it in human english for you.

I guess I told what EXACTLY HP told you

Good luck

Stefan

[Patrick Wessel]

Stefan,
The TLB (transaction lookaside buffer) is responsible to convert virtual addresses to physical addresses and the other way around. The pdir (page directory) of the loaded by the IO drivers. Than means, if you have a problem with a bad pdir entry you need to check the driver patches. Unfortunately are we not able to nail the driver down, because the U2, who pulled the error, is the runway bus to HSC converter.
We may find some more hints in the io error log information of the HPMC. I can't promise anything, but it's worth a try if you want to post the complete tombstone file