ia64_corehw, lpmc_em, disk_em, dm_corehw eats CPU

[Luc Lecocq]

Hello,
I installed HPUX 11.23 December 2006 on two identical L2000-44.
On one of them, I have these 4 processes eating almost continuously between 10 and 20% of CPU.
This does not occur on the second system.

Anyone can explain ?
Thanks
Regards,

[Torsten.]

Could you please run

swlist |grep -i diag

to ensure your diagnostic is current?

I'm not sure if you did a fresh install with 11.23 or update only. In case of update/patching your diagnostics version may be unpatched and old.
Just to be sure.

[Luc Lecocq]

It says:
Online Diag B.11.23.08.04 HPUX Support Tools Bundle Sep 2006

[Sameer_Nirmal]

Can you post the output of
# ps -ef | grep -E 'stm|resmon|logd|psm'

Do you see any difference in terms of h/w setup ( like external storage array etc ) between these 2 servers , clustered?

These processes are EMC h/w monitors . If they are standalone servers , then you can try stopping/killing these monitors using monconfig. Then stop the online diagnostics using
# /sbin/init.d/diagnostic stop

Check if you see any EMS monitor or STM processes command above. Kill them if there are still there.

Then start the diagnostics using
# /sbin/init.d/diagnostic start

This may start the EMS monitors as well. If you don't see EMS h/w monitors running, start/enable them using monconfig.

[Luc Lecocq]

Hi,

I have observed this a full day after the IFS, however, after the last wee-end, it seems the problem disappear, at least for the moment.

Here is the info you requested:
root 735 1 0 Jul 19 ? 0:00 /usr/sbin/syslogd -D
sfmdb 1918 1 0 Jul 19 ? 0:03 /opt/sfmdb/pgsql/bin/postmaster -i -D /var/opt/sfmdb/pgsql
root 2154 1 0 Jul 19 ? 5:13 /usr/sbin/stm/uut/bin/sys/diagmond
root 2318 1 0 Jul 19 ? 0:00 /etc/opt/resmon/lbin/emsagent
root 2282 1 0 Jul 19 ? 0:00 /opt/hpsmh/lbin/smhstartd
root 2899 1 0 Jul 19 ? 97:51 /usr/sbin/stm/uut/bin/tools/monitor/ia64_corehw
root 2554 1 0 Jul 19 ? 0:20 /etc/opt/resmon/lbin/p_client
root 2747 2154 0 Jul 19 ? 0:08 diaglogd
root 2735 1 0 Jul 19 ? 94:37 /usr/sbin/stm/uut/bin/tools/monitor/disk_em
root 2748 2154 0 Jul 19 ? 0:01 memlogd
root 2746 2154 0 Jul 19 ? 1:09 cclogd
root 2749 2154 0 Jul 19 ? 3:47 psmctd
root 2809 1 0 Jul 19 ? 0:11 /usr/sbin/stm/uut/bin/tools/monitor/dm_memory
root 2789 1 0 Jul 19 ? 97:57 /usr/sbin/stm/uut/bin/tools/monitor/dm_core_hw
root 2938 1 0 Jul 19 ? 0:01 /usr/sbin/stm/uut/bin/tools/monitor/sysstat_em
root 2875 1 0 Jul 19 ? 0:12 /usr/sbin/stm/uut/bin/tools/monitor/fpl_em
root 4276 1 0 Jul 19 ? 0:00 /usr/sbin/stm/uut/bin/tools/monitor/msamon
root 2928 1 0 Jul 19 ? 98:51 /usr/sbin/stm/uut/bin/tools/monitor/lpmc_em
root 14230 1620 0 09:30:39 ? 0:00 /etc/opt/resmon/lbin/registrar
root 14611 14450 1 09:33:30 ttyp4 0:00 grep -E stm|resmon|logd|psm

[Andrew Merritt_2]

Hi Luc,

Are there any EMS notifications being reported in /var/opt/resmon/log/event.log?

The CPU usage for those monitors looks unusually high. I would recommend opening a support call with HP for them to investigate further. Have there been any resource problems on the system?

Andrew

[Luc Lecocq]

I looked in event.log, and did not see anything, in fact it's almost empty.

And as I said in my previous post, this behaviour last for one day and disappeared since then.

I don't know what to think about this.
If I am not able to reproduce it, it will be hard to investigate.