topic Re: LPMC errors in Operating System - HP-UX

LPMC errors

Prashant Zanwar_4 — Thu, 07 Oct 2004 11:28:27 GMT

Hi,
I have bunch of LPMC errors as attached in file here. I really wonder what do they mean, is it something problem in the system. What causes these messages.
Few are like this:

Monitor Arg[0] = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Dictionary Name = "/system/cpu/lpmc"
Description = "LPMCs for the Processor"
Monitor Name = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Monitor Arg[0] = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Dictionary Name = "/system/cpu/lpmc"
Description = "LPMCs for the Processor"
Monitor Name = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Monitor Arg[0] = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Dictionary Name = "/system/cpu/lpmc"
Description = "LPMCs for the Processor"
Monitor Name = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Monitor Arg[0] = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Dictionary Name = "/system/cpu/lpmc"
Description = "LPMCs for the Processor"
Monitor Name = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Monitor Arg[0] = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Dictionary Name = "/system/cpu/lpmc"
Description = "LPMCs for the Processor"
Monitor Name = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"
Monitor Arg[0] = "/usr/sbin/stm/uut/bin/tools/monitor/lpmc_em"

Has any one come across these. Please let me how to resolve these.

Thanks
Prashant

Re: LPMC errors

A. Clay Stephenson — Thu, 07 Oct 2004 11:38:49 GMT

Low Priority Machine Check. A very typical cause of these is a problem in one of the on-board caches (e.g. the instruction cache). The box recogized the error and corrected it and a small number of these (1 or so a week) is normal. Cosmic rays and other background radiation can actually cause a few of these but if you are seeing a few (~ >5/wk) then it is an indication that you have a hardware problem. Almost always the fix is a processor board replacement.

Re: LPMC errors

Bill Hassell — Thu, 07 Oct 2004 11:43:18 GMT

LPMC = Low Priority Machine Check which means that a correctable error has occurred. Most often, this is a correctable parity error in RAM but it may be something else. It is a hardware error that has not yet caused a failure. Run the cstm tool to see what is happening:

# echo "selclass qualifier cpu;info;wait;infolog" | /usr/sbin/cstm

and

# echo "selclass qualifier memory;info;wait;infolog" | /usr/sbin/cstm

That should give you a summary of what is about to fail. In the second report (memory) make sure that your PDT (Page Deallocation Table) is empty or almost empty. If it has many entries, memory is going bad.

Re: LPMC errors

Prashant Zanwar_4 — Thu, 07 Oct 2004 12:05:36 GMT

Actually these erros are quite old. I mean last occured on June 24 around. And I am seeing it from 1999 onwards.

Server model is 9000/800/N4000-4.

I am working on this setup just for last few weeks.

I am attached here the memory & CPU information logfile by STM. Further advise is most welcome and appreciated.

Thanks
Prashant

Re: LPMC errors

Prashant Zanwar_4 — Thu, 07 Oct 2004 12:05:59 GMT

2nd file

Re: LPMC errors

Kent Ostby — Thu, 07 Oct 2004 12:10:48 GMT

Have you had any CPUs replaced on this box ?

The errors would seem to be indicating that CPU with a hard physical address of 0xfffffffffed25000.

You would need to have HP hardware troubleshoot this and replace the correct CPU board.

Re: LPMC errors

Prashant Zanwar_4 — Thu, 07 Oct 2004 12:12:29 GMT

Thanks a lot for above replies.

Sorry but these errors are quite old. Last was June 24 2003 seems. Does this still have significance if there are no errors now.

Is there any way to know if any CPU board was changed since then and or a memory borard.

Thanks
Prashant

Re: LPMC errors

Prashant Zanwar_4 — Thu, 07 Oct 2004 12:21:52 GMT

Kent how did you know that hardware physical address # might have a problem. Please reply.

Thanks
Prashant

Re: LPMC errors

Prashant Zanwar_4 — Thu, 07 Oct 2004 14:06:01 GMT

Any further light on this please.
Kent can you please tell me how did you analyse the hardware path processor # might have problems..

Thanks
Prashant

Re: LPMC errors

Sundar_7 — Thu, 07 Oct 2004 19:03:54 GMT

hmm..I dont see any LPMC errors at all.. The text in your original questions are only some kind of definitons.

Refer /var/opt/resmon/log/event.log* files

If there were any LPMC errors, you would see them in these files.

Re: LPMC errors

Prashant Zanwar_4 — Fri, 08 Oct 2004 09:25:32 GMT

OK...Good to hear from you there...I will discard the errors then.

Thanks
Prashant

Re: LPMC errors

Prashant Zanwar_4 — Fri, 08 Oct 2004 09:42:08 GMT

Thanks