memory issue

[Ravikant]

Hi All,

Every couple of mintutes, I can see lots of chmod, chown, chgrp proceses running on HP-UX server died state.

This is the output from glance tool..

chown 18731 root 0.4 1 1.2 47.5mb 81.0mb died
chown 18716 root 0.4 1 1.2 47.5mb 81.0mb died
chown 18735 root 0.4 1 1.2 47.5mb 81.0mb died
chmod 18705 root 0.4 1 1.0 47.5mb 81.0mb died
chown 18720 root 0.4 1 1.2 47.5mb 80.9mb died
bdf 18753 root 0.4 1 1.0 47.5mb 81.0mb died
chmod 18732 root 0.4 1 1.0 47.5mb 81.0mb died
chmod 18714 root 0.4 1 1.0 47.5mb 81.0mb died
chmod 18726 root 0.4 1 1.0 47.5mb 81.0mb died
df 18736 root 0.4 1 1.0 47.5mb 81.0mb died
chown 18725 root 0.4 1 1.2 47.5mb 81.0mb died
chown 18713 root 0.4 1 1.2 47.5mb 80.9mb died
chown 18706 root 0.4 1 1.2 47.5mb 81.0mb died

It is happening very frequently. I cheked any cron jobs running on the server for the same but nothing is present on the same..

Pls help..

Regards,
Rav

[Mel Burslan]

Most probably, an agent of some third party application is responsible for running this many processes at one time. It sounds like this is some sort of a monitoring agent to me but it is just a guess. Try to find the parent process id of the chown processes from the output of command 'ps -ef' and see what process is launching them.

[Dennis Handly]

Is anything using "swverify -F" or using swinstall?

As Mel said, using ps(1) may give you a clue what's going on.

[James R. Ferguson]

Hi Rav:

Moreover, whatever is running is lousy, needlessly (probably) resource-intensive code.

Both 'chown' and 'chmod' allow a list of files to be specified. In that way, _one_ process can suffice to change many, many files. This will be far faster and less resource-intensive than the repetitive calls.

Regards!

...JRF...

[Steven E. Protter]

Shalom Rav,

A script or application is trying to run chown and chmod processes and they are not terminating normally.

This is almost always bad code.

SEP

[Dennis Handly]

>SEP: and they are not terminating normally.

I'm not sure how you know this? All processes die, some abnormally.

[Ravikant]

Thanks everyone..Got the details..one of the monitoring tool running those processes..

We are taking up with monitoring team....

P.S. I don't see the point tab enable to provide the points to u..:-(

regds.
Rav

[Dennis Handly]

>I don't see the point tab enable to provide the points to you..:-(

http://forums.itrc.hp.com/service/forums/helptips.do?#33

There is a submit points button on the bottom left. There are drop down points menus for EACH reply.