runaway dm_core_hw process

[Peter Heinemann]

On two occassions we have had the dm_core_hw process run away, consuming 100% of a CPU and eventually causing a swap space consumption alert.

A kill -9 is not effective. Previously we had to reboot to clear. Has anyone seen this before? Is there a patch/workaround?

[Peter Heinemann]

Some additional background info:

hp-ux 11.11
rp8400

[Sebastian Galeski_1]

hi
Try to use
As root:

/etc/opt/resmon/lbin/monconfig
(K)ill monitoring
(Q)quit

If You install last OnlineDiag your problem disapear

hope it help You

[PIYUSH D. PATEL]

Hi,

Go thro this patch.....

http://support1.itrc.hp.com/service/cki/patchDocDisplay.do?patchId=PHKL_26405

You may need to apply the patches since the dm_core_hw is a common problem.

You can also extensively search the patch site with the OS which you are using 11.0 or 11i.

pIYUSH

[harry d brown jr]

Peter,

Check your stm logs, you are probably experiencing issues with your server.

live free or die
harry

[S.K. Chan · Accepted solution]

Seen it on my L2000 (11i), after installing all necessary patches does not help and the fix is to get HP to install the latest firmware revision for you. Find out what's your current PDC firmware by using STM.
# cstm
cstm> map
cstm> sel dev
cstm> info
cstm> infolog

Look for the line "PDC Firmware Rev".

[Martin Burnett_2]

Hello Peter,

First we need to know the version of EMS and STM installed. This is diplayed when /etc/opt/resmon/lbin/monconfig first starts up.

Second check the log files in the directory /var/opt/resmon/log/ for error messages. Particularly client.log, api.log and registrar.log.

Third check that all of the required patches are installed for the version of EMS/STM you are using. You can find the required ones at this web site:

http://www.docs.hp.com/hpux/onlinedocs/diag/st/st_read.htm

They are listed by date of release for the seperate versions of EMS/STM. To resolve your versions release date use the table of EMS versions located at:

http://www.docs.hp.com/hpux/onlinedocs/diag/stm/stm_upd.htm#table

Last did this behavuoir just start or were there any other changes made? Did it work properly before?

Thanks for participating in the forums,

Martin
Chaos reigns within. Reflect, repent, and reboot. Order shall return.

[Peter Heinemann]

...and everyone shares in the points!

I logged the problem with HP CSS, just to cover all bases.

Sebastian: I was able to stop the process using your approach; and CSS support said that was the offical workaround.

Pujush: The patch pointed to similar (but worse) behavior. We were in fact seeing the process run away on both partitions, although no panic (yet) which the patch corrects.

Harry/Martin: the api.log pointed to a problem with a crossbar controller chip.

SK: turns out to be known pdc firmware problem to be fixed in the next release.

Thanks, all.