STM and Cluster

[David_711]

Hi, i have a problem with stm on two servers rp3410 with service guard installed. I always watch zombies proceses on both servers and the proceses are stm proceses.
Somebody know if there is a incompatibility problem with service guard and STM.
I upgrade the STM to the june version and the same process are present.
Please help.
Thanks.

[Mel Burslan]

David,

For what it is worth, I have only one rp3410 server running B.11.11.14.15 version of stm and I keep seeing a defunct process, with a ppid of a process which is a corehw monitor tool.

I am not sure what the root cause of this and never had a strong enough urge to investigate as it is not causing me any performance issues right now.

Once I contacted HP about it and their suggestion was to upgrade STM but it never got done.

You are not alone :)

[Andrew Merritt_2]

Hi David,
This is normal behaviour for the ia64_corehw monitor, which I think, if you look at the 'ps -ef' output, you will find is the parent process of the process.

The monitor forks a child every minute. The child processes some of the log files, and then exits. The parent reaps the child by doing a 'wait' call every minute. Therefore, if the log processing is done quickly (which is normally the case), you will see a child of ia64_corehw most of the time. Note, however, that the process id of this is changing every minute.

I'd agree this may not be the most efficient way of doing things, but it should not cause any problems, and is not using any system resources beyond a process slot. There are currently no plans to change this behaviour.

Andrew