Unable to kill process....

[Sarjerao]

Agent run from a startup script /sbin/rc3.d starts but when attempted to shut it or kill it, the processes didn’t killed.
The agent processes that stay alive cannot be killed even with kill -9.

But insteed to startup script, after system boot, if agent start then I can stop/kill the process.


Is there any other way to automate the process? Thanks in advance.

[James R. Ferguson]

Hi:

If a 'kill -9' doesn't terminate a process it's because the process is waiting on an evert (often an I/O) that until satisfied cannot allow the process to resume.

For a signal to be acted upon, the process must be running.

Regards!

...JRF...

[Sarjerao]

IS there any other way to stop this process. Because due to this I unable to start or stop it.

[Jannik]

Normally it is do to a missing mount point and normally of the type nfs.
What does your dmesg tell you. If it is a missing nfs mount point you can start the nfs server again on the server site of the nfs. The you should be able to stop the processes.

The mount command will give you the needed information.

Does the bdf command hang?

You could reboot you system but first edit the files that do the mount of the nfs (/etc/fstab, cluster conf or automount in the etc.

It could be a failed mountpoint the dmesg and /var/adm/syslog/syslog.log should give some information. and you can use the vgdisplay -v vgXX to get information on the volumegroups.

[Laurent Menase]

Hi,


If kill -9 doesn't work, it means it is an application hang in the kernel.
If it is not related to a hard NFS mount hang,
you should take a 'TOC' of the system, and contact hp support.