Operating System - HP-UX
1834956 Members
2077 Online
110071 Solutions
New Discussion

How to get rid of defunct process name of ia64_corehw

 

How to get rid of defunct process name of ia64_corehw

Hi~
My clinet want to eliminate the defunct process.
The result of searching the process.

cschp1:/>ps -ef|grep 2104
root 2104 1 0 May 21 ? 1:35 /usr/sbin/stm/uut/bin/tools/monitor/ia64_corehw
root 24610 2104 0 11:02:24 ? 0:00
root 25137 21863 1 11:03:20 pts/th 0:00 grep 2104

The STM version is DEC 2004.

Thanks in advance for your time.
JS.Chae
4 REPLIES 4
Amit Agarwal_1
Trusted Contributor

Re: How to get rid of defunct process name of ia64_corehw

What version of HP-UX are you using?

On 11i v2, init(1M) takes care of cleaning zombie process at periodic intervals. So your client need not worry about it.

Having said that, I would suggest your client to investigate the actual cause of these zombie processes. Who is creating them, and why are they getting created. Its in the best interest to rectify the app that is creating zombie. All it needs is a wait*() systemcall to be placed at right location in the source code.
Joseph Loo
Honored Contributor

Re: How to get rid of defunct process name of ia64_corehw

hi,

is a process that has exited and has a parent, but has not yet been waited for by the parent. this is also called zombie process.

if u r want to stop the process, u may use:

# /etc/opt/resmon/lbin/monconfig
> select (K)ill to disable
# ps -ef|grep -i stm
to check if STM monitoring is still on

# /etc/opt/resmon/lbin/monconfig
> select (E)nable to enable


regards.
what you do not see does not mean you should not believe

Re: How to get rid of defunct process name of ia64_corehw

Thank you.
HP-UX version is 11.11 and The Server is Rp3410
Andrew Merritt_2
Honored Contributor

Re: How to get rid of defunct process name of ia64_corehw

Hi,

Joseph's definition is an accurate description of the situation; the defunct process is not using any resources except for a slot in the process table.

This is a feature of the way the ia64_corehw monitor is implemented. If you look carefully, you'll notice the defunct child is never more than a minute old. A child is spawned to process one of the log files. It exits when it reaches the end of the file. The parent process issues a wait every minute, so will reap the dead child process at that time. If it takes more than a minute to process the log file, the parent will detect that the child still lives and doesn't start a new one.

I wouldn't have done it that way myself, but there's no actual defect in the way it works.

What problems is the presence of the process causing?

(In earlier releases of the OnlineDiags there was a bug that meant that multiple defunct children existed in some cases, and for longer than a minute, but that is fixed in the version you have.)

Andrew