HPUX11.0 Tombstone - server crash - any ideas?

[Roland Rebstock_1]

All, this server has crashed twice in 2 weeks generating a tombstone file only. Can someone check from the tombstone file attached what is wrong. We called our support vendor but they have not had any luck finding out what is wrong. I looked to cstm and found about 250 errors on on memory dimm but im unsure if they are old.

[Patrick Wallek]

Anything in /etc/shutdownlog or /var/adm/crash or /var/adm/syslog/OLDsyslog.log?

[Alex Glennie]

can you let us know the model here ... it's a PA Risc system i know ...

I maybe able to help if a newer system

[Alex Glennie]

fyi if this is an rp34xx,rp44xx or a rp74xx as I suspect it maybe the decode of the ts99 file doesn't give any clues ....

perhaps out of date f/w or diags ? I'd suggest
also looking at /etc/shutdown.log , the OLDsyslog.log and seeing if a /var/adm/crash/crash.## was also created. if so call HP support or get your support vendor and ask them to ?

[Sameer_Nirmal]

Roland,

Looking at the tombstone file, it seems that processor 2 doesn't have valid timestamp. Saying that it didn't respond to the HPMC occured. It might be dead.
Did you check anytime the number of processors being used in between the two crashes?

[Andrew Rutter]

hi roland,

from the tombstone file it loks like you have an N4000 system with 6 processors?

It would also be interesting to the latest 6-10 error logs in the GSP.

If you have alot of memory errors then this could be a major factor, especially if there double bit errors. single bit errors the system should be able to recover from.

You also have no timestamp for proc 2 as already mentioned, this looks like it could be an issue aswell.

You should run tests on the cpu's and memory in STM and gather the info from infolog tool. It would be interesting to see if proc 2 has been deconfigured?

Andy

[Andrew Rutter]

roland,

did you find the cause of your problem? as you havent come back yet with any more information, or to let us know the final outcome.

If its resolved please post the answer as it may help others and close the thread
Might be a good idea to assign a few points where required aswell, to this and some other threads you havent done so?

Andy