psmctd process using up 100% cpu time!!!

[Rajasekhar Mamudur]

Hi All,

I hope this is the correct forum for this quastion!

I have the psmctd process taking up 100% of cpu time of 1/12 cpus on a HP 9000/800 v2600 server. HP-UX version is 11.11. My questions are:

1. I want to know why it is taking up 100% time! I just know that it monitors peripheral devices. I want to know if I can run any diagnostics and see what exactly the issue is!

2. Sample output from /etc/opt/resmon/log/client.log
-------------------Start Event--------------------
User event occurred at Mon Aug 4 10:58:05.984443 2003
Process ID: 21983 (/etc/opt/resmon/lbin/startmon_client) Log Level: Error
Monitor request was not accepted
-------------------End Event----------------------

-------------------Start Event--------------------
User event occurred at Mon Aug 4 10:58:05.987595 2003
Process ID: 21983 (/etc/opt/resmon/lbin/startmon_client) Log Level: Error
Error in receive_monitor_reply function.
: Internal error: error code = 1878986968: Socket is already connected
-------------------End Event-------------------


3. What other log files give additional info? I would like to investigate the issue, find the cause and fix it!

4. Will retsrting the process resolve the issue? If so how can I do it? How will this affect psmmon process?

Any help is greatly appreciated!

TIA,

Raj.

[harry d brown jr]

using "lsof" determine what has ports 49212, 49312 and 49313 open.

lsof | grep -e 49212 -e 49312 -e 49313
psmctd 2227 root 1u inet 0x492b7d00 0t67925 TCP vpart1.paychex.com:49212->vpart1.paychex.com:registrar (ESTABLISHED)
psmctd 2227 root 2u inet 0x49027800 0t0 TCP *:49312 (LISTEN)
psmctd 2227 root 4u inet 0x49027e00 0t0 TCP *:49313 (LISTEN)
registrar 2234 root 0u inet 0x4994d1c0 0t67925 TCP vpart1.paychex.com:registrar->vpart1.paychex.com:49212 (ESTABLISHED)
registrar 2234 root 1u inet 0x4994d1c0 0t67925 TCP vpart1.paychex.com:registrar->vpart1.paychex.com:49212 (ESTABLISHED)
lsof: WARNING: /.lsof_vpart1 was updated.
#


if you don't have "lsof" get it from here: http://hpux.cs.utah.edu/hppd/hpux/Sysadmin/lsof-4.67/

live free or die
harry

[Ron Barnett · Accepted solution]

Also try this:
if problem with psmctd deamon ( not running ) do this :

cd /var/stm/data
mv psm_data psm_data.old
touch psm_data
as root : /etc/opt/resmon/lbin/monconfig ...then k for kill
Stop diagnostics:
/sbin/init.d/diagnostic stop ( wait untill all prcesses are killed) this will ta
ke a while
/sbin/init.d/diagnostic start ( wait unit all processes are started) this will t
ake a while.
as root : /etc/opt/resmon/lbin/monconfig ...then e for enable

[Elena Leontieva]

Raj,

Take a look at this document:
Date: 5/28/03
Document description: psmctd is using alot of CPU.
Document id: KBRC00012536

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000068695470

Elena.

[Rajasekhar Mamudur]

Hi All,

Thank you Harry, Ron and Elena.

Harry: I was able to install lsof but, I got an error when ran it. I have not been able spend time on it to get it going. I will fix it soon.

Ron:Thank you very much, I was able to stop it and start the diagnostic and the related processes.

Elena: Thanks for that link. That did not work for me. The file was present for me.

I read somewhere that booting the server twice in quick succession may cause this problem. This did happen the previous weekend, but we did not face the issue until 3 days back.

What is the relevant document that describes the diagnostics? Is it the EMS guide?

Thank you.

Raj.