ia64_corehw

[Sachgebiet 341]

Hi all,

on HP-UX 11,11 we have an process "ia64_corehw" that loops. Can we kill this process manually with a kill -9 and the next question is, can we stop this process from starting again because we have this problem on 2 machines.

Did the process fpl_em that sometimes loops also will be stopped, if we can stop the other process?

Thanks for help

Andi

[Stf]

Hello,

It seems to be an EMS hardware monitor for ia64.

Look at this:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?admit=716493758+1103100830078+28353475&threadId=7317

I think you should not kil it...

Stf ;-)

[Sachgebiet 341]

Hi,

Thanks for answer, but in this thread there stands nothing about stopping this process. This process loops on a productive machine. It would be nice if we can stop this process. It takes sometimes 50% from our cpu on a 7420.

Greetings Andi

[Stf]

Ok if you really want to stop it I think you should try:

# vi /etc/rc.config.d/diagnostic
DIAGNOSTIC=0
then
# /sbin/init.d/diagnostic stop

But I think this is dangerous, because you probably won't be able to detect failure in the future...

Stf ;-)

[Stf]

It seems this patch would be interesting for you...

http://www.informatik.uni-frankfurt.de/RBI/hp-service/patches/s700_800/11.X/PHSS_31985.txt

look at symptoms

Stf ;-)

[Andrew Merritt_2]

What version of OnlineDiags do you have installed? And what exactly are you seeing that indicates it is looping?

Yes, you can try killing it with -9; a new copy of the monitor should then be spawned, so if it was some transient problem that might cure it, but in my experience that's unlikely.

Is there anything logged in /etc/opt/resmon/log/api.log or client.log?

If it's not a recent version of OnlineDiags, consider updating to the latest version, and applying the latest patch.

Andrew

[Andrew Merritt_2]

If it's just that the accumulated CPU usage seems high, that is a known feature of these two monitors. CPU usage will be higher than for the other HW monitors because of the way they work. This is partly dependent on the hardware configurartion.

There is an enhancement request open for fpl_em - 8606384740/JAGaf44894

Andrew

[Girish_17]

Andi,

I am facing similar problem but i am not worried about the performance but the defunct processes it is creating. I always sea process whose ppid is of ia64_corehw. Is that the case with you?

I am also working around the same problem and time being disabled diagnostics (which is not at all recommended).
If you want to do so you will have to edit three files in /etc/rc.config.d
ems,emsagtconf and diagnostic. Edit this file and replace =1 with =0. And then you will have to reboot your server (ofcourse if you are able to reboot).

Well now i don't see any defunct process and you may not see them looping and utilizing processor heavily.

Thought it could help you.
Regards,

[Sanjay_6]

Hi,

There are two things.

defunct child processes it seems for "ia64_corehw" is normal.

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000074246192

The doc id is 8606344112.

The next problem is that "ia64_corehw" has memory leak problem. expected solution from hp in dec 2004 release of HWE

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000076588987

The doc id is 8606380199.

hope this helps.

Regds