Glance display too many died process !!!

[Bill Hassell]

You have massive runaway processes and they appear to be in an endless loop. The processes marked as DIED will not appear again (the process ID keeps changing). The processes start to run and get killed by some other process. Your system is very sick and unless you know the process(es) that are causing this condition, you'll have to reboot into single user mode and stop any scripts or programs that were added before this problem started occurring. It appears that this may be part of a cluster so if possible, failover this node to the other and troubleshoot the problem.

[HP-UX_Ali]

Thanks Bills.. appreciate your quick response and advice...

You are right this is env is 4 node metro-rac cluster and the issue arises in one of the DR site (2nd node) which is not in much use but i wonder why this is happening and after some troubleshoot, i noticed their is no such script or programms which is killing any process but one thing i suspect does it due to crs process ?... , i want to troubleshoot this without reboot....

Once again thanks for your support and time.

Thanks & Regards
Ali....

[Steven E. Protter]

Shalom,

This is a problem. These processes are hung and have not released their memory.

They are related to serviceguard.

SEP

[Dennis Handly]

>dennis I didn't got your point on this.

As Bill said, you seem to have continual cmclcond processes that just keep coming back.
And more kill & sleep.

And of course, while they are alive, they take up memory and CPU resources.

>I want to troubleshoot this without reboot.

I'm not sure how successful you'll be.
You might start with a hierarchical process tree:
UNIX95=EXTENDED_PS ps -ef -H