Problems to kill processes

[Manuales]

Hi !!!
Every day its run a process into the server, this file remove files .. it is every hour, right now i can see like 20 of this one running, also i have a file system full, i need to kill them but i can't do it .. how can i kill them ?
i'm running:
kill -9 user-id (with root) but it can not do it !!!

Thanks ...

[James R. Ferguson · Accepted solution]

Hi Manuales:

If you are running as 'root' and a 'kill -9' doesn't kill a process, then the process is in an un-killable state -- waiting on a signal, I/O completion or running in kernel code.

A full filesystem can cripple a server. Try manually removing a file, if you can, to regain space.

Regards!

...JRF...

[A. Clay Stephenson]

First, you need to improve your cron'ed script so that multiple instances of the script do not run. The most straightforward way is to test for the existence of a lock file and if found exit; otherwise, you create the lock file and execute. You then remove the lock file atr the end of this script so that the next time the script is executed no lock file is found.

Kill is perhaps the worst named system call and command in UNIX. It should really be called something like "sendsignal" --- that's all a kill actually is. If a process (like yours) is waiting on a higher-priority event such as an i/o operation, it will never respond to a signal until the higher priority event is satisfied. If the system didn't work like this the things like filesystem corruption (because the signals stopped incomplete actions) would be very common.

Almost certainly, you are going to have to reboot to fix your problem but if you fix the script as I mentioned earlier, your other problems willl probably not arise again.

[Manuales]

thanks...
also, with dmesg command its reports:

DIAGNOSTIC SYSTEM WARNING:
The diagnostic logging facility is no longer receiving excessive
errors from the I/O subsystem. 7 I/O error entries were lost.
DIAGNOSTIC SYSTEM WARNING:
The diagnostic logging facility has started receiving excessive
errors from the I/O subsystem. I/O error entries will be lost
until the cause of the excessive I/O logging is corrected.
If the diaglogd daemon is not active, use the Daemon Startup command
in stm to start it.
If the diaglogd daemon is active, use the logtool utility in stm
to determine which I/O subsystem is logging excessive errors.

what does last one mean?

Thanks.

[James R. Ferguson]

Hi (again) Manuales:

I would expect 'dmesg' to contain the *critical* alerts. HP-UX is trying to help you find why it feels so "sick".

..."If the diaglogd daemon is active, use the logtool utility in stm to determine which I/O subsystem is logging excessive errors"

...means that you could run '(c)stm' [ the Support Tools Manager ] to obtain more detailed diagostic information about the execessive I/O errors. You don't need to do this, since you already know the reason for the alerts!

Regards!

...JRF...

[A. Clay Stephenson]

You have a failing disk or controller (or if you are fortunate something as simple as a missing or bad terminator). You should see plenty of messages in /var/adm/syslog/syslog.log that will indicate the failing device and there is a good change that ioscan -fn will display NO_HW.