Process ia64_corehw defunct

[affin]

Dear Guru's,


on our rp3410 box we found a zombie process as below:

root 3190 1 0 18:06:29 ? 0:00 /usr/sbin/stm/uut/bin/tools/monitor/ia64_corehw

we tried to reboot the system but the event comes again.

Any suggestion?


Many Thanks,
Alessandro.

[Patrick Wallek]

That appears to be one of the diagnostics monitors. Though why an ia64 monitor is comeing up on a PA-RISC machine is something I'm not sure about.

I would query HP about it. But as long as it isn't causing any problems or eating CPU or Memory I wouldn't worry about it.

[RAC_1]

This could be EMS issue, Are you up to date on STM/EMS patches??

Anil

[Tony Scully_2]

I believe this was supposed to be fixed in a December 04 Support Tools Manager patch, but I cannot find one. The defunct process (zombie) is actually cleaned up and then re-created (check the PID, it will change) and does not cause any problems (uses no resource) but it is untidy.

Tony

[Dave Hutton]

I actually opened a ticket with HP about this. I would say a handful of servers have this process defuncted.
I was never really given a good answer to why.
Where I really noticed this was after I ignited the server from a different server.
I was basically told to ignore it. EMS/ISEE/STM all seem to still function fine.

[Cheryl Griffin]

Patrick - ia64_corehw monitors both IA and PA hardware

The monitor does respawn a defunct process. This is by design. The March 2004 release of Diags for HP-UX 11.23, and the December 2004 release of Diags for HP-UX 11.11 contains a fix that is supposed to optimize the process.

If you monitor the pid of ia64_corehw you'll see the pid actually changes. The monitor is the parent process, which is waiting for a child process to finish. During the wait time the process will appear as defunct.

[Patrick Wallek]

Cheryl,

Good info. Apparently the process name changed for 11.23? I am used to seeing dm_corehw for my 11.0 and 11.11 machines. I just looked at my 11.23 and I see the ia64_corehw now.

Since it started with ia64, and no HP-UX version was given, I figured there was something strange going on.

Live and learn......

[Andrew Merritt_2]

Cheryl has got close to what is actually happening.

The parent ia64_corehw periodically (every minute) forks a child to process one of the log files. The child exits as soon as it has finished doing so (which is usually very soon after starting, as there are not many log entries added). The parent process only checks for the child exiting every minute, so until that time the child shows as '<defunct>'. (If the child was still alive, processing the log file, then a new one would not be forked.)

This is known behaviour, and is not optimal, but it is not incorrect behaviour. It is not having any effect on the system beyond the overhead of forking the child process and dealing with its cleanup each time.

There are currently no plans to change this.

If it is causing a problem beyond being unexpected, please give details either here or via HP support.

Andrew Merritt