C8000, ECC RAM and Log files

[Bill Calver]

We have a C8000 that we have been told contains all ECC RAM (8x4GB dimms). When running simulations we are experiencing various memory failures (memory fault, bus error, etc) while reaching only about 2.7GB. Tombstone files don't reveal anything (latest Timestamp is Aug 17, 2006). ts99 is attached.

Will ECC RAM errors write msgs to tombstone files or will they write to some other log files?

Thx, Bill

[siva0123]

Bill,

Are you facing this error still recently ..

If not there is no use in investigating a ts99 file which has the August 17 Timestamp.


But any how if there is a hardware error , if you have EMS , then the logs should be in /var/opt/resmon/log/event.log.

Is there any errors in dmesg and syslog.

Also any warning or errors in the console logs.

Run a cstm diagnosis for the memory module which may give some clue.

Thanks,
Siva

[Bill Calver]

Hi Siva,

We still have the problem (as of last week) and actually it started after August.

There is no /var/opt/resmon/log/event.log file.

dmesg does report a "file system file data read error" on Nov 8 which was around the time of the latest crash. This sounds like a disk space full msg to me but even on the disk that is 97% full we have 6.5GB free.

STM memory error log is empty.

Didn't see any warning or errors in the console logs.

Bill

p.s. This may "skew the results" except for previously-written log files but we had two of the dimms replaced yesterday by the vendor. He wasn't sure which boards may be failing and that's what prompted my question about ECC error logging.

[siva0123]

Bill,

If that is the case , i'm not sure whether your DIMM's are loaded properly in the memory board . I beleive these DIMM's has to loaded in pair for better performances.
Just a thought though im not sure .

Also was there any paricular change done to the machine or the aplication around august?

And when there was a HPMC in August was HP contacted and was there any diagonosis doen on that ts99 file as i see there are some DIMM errors logged. As HP has a internal tool to diagnoise ts files they should be able to let you know what happened exactly.

Thanks,
Siva

[Bill Calver]

I think they do have to be loaded in pairs and that's why the tech installed two new dimms.

His plan, if/when it fails, is to replace the new mem in slots 4A/B with old mem in slots 3A/B and work his way down (since 4GB dimms are $$$$). This could take a while...

FYI, we received the machine in June and at that time there were errors on boot up. The tech reseated all the memory then and it booted up. We haven't tried to push the machine this hard until now.

Otherwise, there have been no changes to the machine and HP was not contacted.

Bill