Zombie process from ia64_corehw process

[Tony Scully_3]

Hello,

I have an RP8420, running HP-UX 11.11 December 2003 Enterprise OE, patched up to June 2004 (Online Diags, HW enablement and GOLDQPK). I keep getting a zombie'd process, which is a child of the ia64_corehw process, like this:

root 2127 1 0 11:18:29 ? 0:00 /usr/sbin/stm/uut/bin/tools/monitor/ia64_corehw
root 4301 2127 0 13:24:07 ? 0:00

A while later the zombie will have a new PID, so it must be getting cleaned up and recreated:

root 2127 1 0 11:18:29 ? 0:00 /usr/sbin/stm/uut/bin/tools/monitor/ia64_corehw
root 4432 2127 0 13:52:38 ? 0:00

While this is not causing any 'problems' I don't like it.

Can I:

1. Stop running the ia64 process - as this is a PA-RISC box anyway, and I have a dm_core_hw process running as well?
2. Fix it another way?

TIA

Tony.

[Cheryl Griffin]

Tony,
This particular zombie is not a real zombie but how the ia64_corehw process regens itself. If you monitor it, the zombie will be replaced periodically by a new process which is also a zombie.

Look for a patch....

SR 8606344112
Source code to minimize defunct processes. Expected release date: December 2004 OnlineDiag for HP-UX 11.11.

[Tony Scully_3]

Cheryl,

Thanks for that - I could see that the process was being respawned, so I was a bit baffled as to why it was appearing as a zombie.

Do you know what the ia64_corehw process monitor on a PA-RISC architecture box?

I'll keep an eye out for the patch, but whether anyone will let me apply it may be a different problem!

Cheers,

TOny.

[Cheryl Griffin]

You'll be able to apply the patch (or the updated Diagnostics) on an 11.11 PA machine.

ia64_corehw monitors both PA and IA hardware, even though the name appears to be IA only.

According to http://www.docs.hp.com/hpux/onlinedocs/diag/ems/emd_ia64core.htm it monitors temperature, fan and power.

[Tony Scully_3]

Cheryl,

Thanks - I'll leave this running and wait for the patch.

Cheers,

Tony.

[Andrew Merritt_2 · Accepted solution]

PHSS_31113 is already available for A.43.00 (HWE0312), and PHSS_31114 for A.45.00 (HWE0406) of OnlineDiags which have the fix for 8606344112. (From what you say, you will need PHSS_31114.)

This won't stop there being a zombie process, but does fix a bug where there were sometimes several of them, and also cleans more frequently.

ia64_corehw runs on IPF systems, and on HP-PA systems with HP-UX 11.11 and OnlineDiags HWE0312/A.43.00 and later.

On IPF, the monitor reports various events from the SEL.

On cellular IPF and HP-PA systems with IPMI, these events are reported to the FPL, not SEL, so are monitored by fpl_em.

On HP-PA systems, the monitor receives PCI error data, and makes it available to the FPL monitor for further processing. This is the main function of ia64_corehw on PA systems.

[Cheryl Griffin]

Andrew is correct, just checked out the patch and it does address 8606344112. So it appears you can apply patch PHSS_31114 if you have STM A.45.00 (June 2004) installed, or install STM December 2004 (when it is released).

[Tony Scully_3]

I'll apply the patch, as these severs are still under my control.

Cheers,

Tony.