System Reboot

[Joan Fernandez]

Hi everyone.

I have a problem with my old HP9000 800/K210 (10.20 HP-UX)Since few days ago, it reboots with a FTL CBFB message on display. In shutdownlog appears the following messages:

19:36 Tue Nov 12 2002. Reboot after panic: trap type 1 (HPMC), pcsq.pcoq = 0.c
2454, isr.ior = 0.0

I 've seen in this forum that this is a hardware problem, but I don't know how to determine wich part of my box is the cause of the problem.

Thanks in advance for your help!
Joan.

[Ken Hubnik_2]

Your best bet would be to call HP but I suspect it is a problem with one of your processors. HPMC I believe stands for HP microcode.

[Robert-Jan Goossens]

Hi Joan,

http://www5.itrc.hp.com/service/cki/search.do?searchString=KCKBCA00000574&mode=id&submit=Search&searchCrit=allwords&docType=Security&docType=Patch&docType=EngineerNotes&docType=BugReports&docType=Hardware&docType=ReferenceMaterials&docType=ThirdParty

Regards,

Robert-Jan.

[Joan Fernandez]

Sorry Robert-Jan, but your search didn't give me any results. Could you post again the document refered?

Thanks.
Joan

[Vincent Farrugia]

Hello,

If there a ts99 log in /var/tombstones directory? If so, please post it here for us to see.

Vince

[Joan Fernandez]

>Hello,
>If there a ts99 log in /var/tombstones directory?If so, please post it here for us to see.
>Vince

This directory doesn't exists. I see some information on Service -> PIM when booting the box, but I don't know how to save this into a file to send you. It's possible?

[David Bell_1]

Joan,

Take a look at this document. It should be noted that HPMC is High Priority Machine Check and it follows that LPMC is Low Priority Machine Check. The HPMC is typically hardware related.

http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062925563

You may also find some useful information on error codes here:

http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062688680

HTH,

Dave

[Robert_Jewell · Accepted solution]

From with the service menu enter PIM and then write down from each processor the chassis codes that are shown. The first will always be cbf0 and then followed by usually four others. You can look up the codes from the links provided above, or post them here and someone can give it a shot for you.

Incidently, to get the ts99 files mentioned you would have to install patch: PHSS_14980. AFter that you wont have to bring the system down to read the logs.

Bob

[Robert-Jan Goossens]

Hi Joan,

Just having dinner, here it is.


HP 9000 K-Class Servers: Accessing HPMC Information


printable version

path: Home

date: 10/5/01
document description: HP 9000 K-Class Servers: Accessing HPMC Information
document id: KCKBCA00000574





You may provide feedback on this document



--------------------------------------------------------------------------------
HP 9000 K-Class Servers: Accessing HPMC Information DocId: KCKBCA00000574 Updated: 10/5/01 9:26:00 AM

PROBLEM
What is an HPMC and how do you access this information on a HP 9000 K-Class Server?
RESOLUTION
An HPMC error condition means that a high priority machine check has occured. An HPMC is an abnormal condition which has compromised the integrity of system processing. A CPU detects the hpmc and halts the system.

The HPMC information is stored in a non-volatile memory location. To access the HPMC information, first reset the system, either by your normal system shutdown/reset procedure, or from the system console:

NOTE: If the operating system is running, you need to shut it down properly first.

Place front panel key switch in service position
Type cntrl b & then type rs
After the system reset, interrupt the autoboot process if it is enabled:

Hit any key when prompted to interrupt
After the interrupt, from the menu presented, go to the service menu, enter the 'pim' command.

Record the pim report information, specifically:

timestamp
chassiscodes
responder/requestor addresses
I/O module error log
memory error log
NOTE: If you have more than one processor in the system, you need to gather this information for each processor.

This pim report information will be helpful to the HP Response Center, in determining the cause of the HPMC


KEYWORDS
HARDWARE, Servers, K-Class

Kind regards,

Robert-Jan.

[Joan Fernandez]

Well, after apply patch, I obtain a ts99 file in last crash. I attach the file in this post.
Anyone could help to interpreter this file?

Thanks!
Joan

PS: I make a mistake on a puntuation, I don't have the problem 100% solved, yet

[Robert-Jan Goossens]

Hi Joan,

Best to call HP-CRC. They now how to read and interpret the tombstone file. All I can say is the ts99 file produces more errors memory, i/o and cpu2 (maybe an i/o board).

Hope it helps,

Kind regards,

Robert-Jan.

[Robert_Jewell]

Your systems suffered an MSI read timeout (indicated by 7d03). This condition will trigger a path error that usually points back to the offending component. In this case CPU0 reported hex 5002 which is a path error. Basically CPU 0 was trying to read into a memory address when it faulted. This caused the report of an MSI read timeout and a path_error by the CPU. You could try reseating the CPU and see if the system crashes again.

There is a chance that this could be software though. There was an old bug that was addressed by some patches, but basically make sure you have the latest critical patches installed (from the last support CD, 12/01.

Good luck!

-Bob