system rebooting & memory error in syslog

[bhavin asokan]

hi,

my V class server is rebooting intermittently. it is giving no error message in syslog.in shutdownlog some times it is giving panic but not for alll reboots.
HP engineer came and replaced the memory module
three days before.but the problem exists.
after changing the memory it is giving following error in syslog.

vmunix:memory module 0x0 has spa shift of 0 , but pdc doesnot support interleave!

can somebody tell me what may be the problem.
machine rebooted today also.but it was unable to save the dump in the crash folder--giving error message 'not able to save the uncomplete dump'.

regds,

[Patrick Wallek]

I would get HP back out there ASAP to look at that machine again! If they have closed the call get it reopened NOW!

It sounds as if one of the RAM modules still has a problem.

[Dave Olker]

Hello Bhavin,

I researched your error message and found the following answer in HP's Knowledge database:
_________________________________________

QUESTION

I have a message in syslog from my V-Class server

Mar 22 10:26:09 unknown vmunix: Memory module 0x0 has spa shift of 0, but PDC does not support interleave!

What does this mean?


ANSWER

Interleave is assigned by POST. After bootup tryes HP-UX again to build inter-leave-groups using the pdc. But SPP_PDC does not support interleave, because POST did it already.

This is a information message only, it is not broken.
_________________________________________

It sounds to me like you can safely ignore that memory error in syslog, as it seems informational only.

Your first order of business seems to be configuring your system to allow a complete memory dump to occur should your box panic again. The following document details exactly how to configure your system for a good memory dump:

http://docs.hp.com/cgi-bin/fsearch/framedisplay?top=/hpux/onlinedocs/5187-2216/5187-2216_top.html&con=/hpux/onlinedocs/5187-2216/00/00/55-con.html&toc=/hpux/onlinedocs/5187-2216/00/00/55-toc.html&searchterms=space%7ccrash%7cdump%7cswap&queryid=20040806-230331

Good luck,

Dave

[Shaikh Imran]

Hi,

Please run the STM and take the information of the memory this could give you some more details.
If possible pls post the details here also.
I have faced the same problem with L-class
& the memory modules were giving single bit error on many DIMM's and they were replaced
and everything was fine.

Also i would advise you to see if you could disable the memory interleaving from the PDC.
if option is available ( No idea)
(Pls take HP's help for this).

Regards,

[Mohanasundaram_1]

Hi Bhavin,

Dave Folker has the best possible solution here. If your systems are indeed V-class then you must be based in Mumbai.

Check with HP Engineers in Mumbai for this problem. The rebooting may be due to some other reason.

With regards,
Mohan.

[bhavin asokan]

hello dave,

if it is not the problem with memory why this error message was not appearing on earlier syslogs.i am having three V class servers - only on this server it is giving error.i am having syslog files for the last 3-4 months and it never appeared on old reboots.

regds,

[Dave Olker]

Hello Bhavin,

I honestly have no idea why the Vclass suddenly started logging these errors when it didn't before. My only guess would be that the CE who replaced your memory boards made some change to the POST or PDT settings during the repair that caused these messages to appear. I recommend you contact the HP support personnel who did the repair and verify if any changes were made during the repair.

Some other things to consider checking yourself, or having an HP support engineer verify:

Are all of your Vclass systems running the same firmware revision? Are the POST/PDC settings identical on all of your Vclass systems? Is there a POST/PDC setting that controls memory interleaving that could be modified to stop this error from printing? (Unfortunately I don't have access to a Vclass system to check this myself.) Is there a newer firmware version for the Vclass that may address POST/PDC issues such as this one?

Whatever you end up finding out during your investigation, please post your findings here as they will undoubtedly help other ITRC forum members.

Regards,

Dave