Hi folks!!
Here is a list of responses I got for my posting.
I have'nt had a chance to test any of these yet because I already
had rebooted my M/C. Next time when I encounter the problem,
I am going to test these.
Thankyou all of you for responding.
Sudheer.
The original posting was:
> I have several processes running on the Sun Server (690MP)
> out of which, a few are eating away a large amount of swap space.
> I tried to kill them, but I can't seem to.
>
> The previous time it happend, I had to reboot the machine.
>
> %ps -axr | grep pmm
>
> 1904 ? D 0:00 /condor/c6/pmm/LASTIP/bin/lastip.exe grade.sol
> 2070 ? DW 0:00 /condor/c6/pmm/LASTIP/bin/lastip.exe nwell.grade
> 2108 ? DW 0:00 /condor/c6/pmm/LASTIP/bin/lastip.exe grade.sol
> 2159 ? DW 0:00 /condor/c6/pmm/LASTIP/bin/lastip.exe nwell_grade.geo
>
> I do kill -9 proc#
>
> Obviously, the processes are non-interruptible and are waiting for I/O.
> Is there a way that I could kill them without rebooting?
The responses are:
---------------------------------------------------------------------------
You can play around in the kernel, but it is somewhat dangerous. For
processes hung in I/O, reboot is generally the answer. Of course,
the real answer is to find out why they are hanging...
l & h,
kev
-----------------------------------------------------------------------------
Unless you can work out what they are stuck on and unstick the device
for the process then the only way to get rid of a D process is to
reboot.
-- Brett Lymn, Computer Systems Administrator, AWA Defence Industries -----------------------------------------------------------------------------I don't really know BUT that looks like it could be running Condor, which is a program to allow (among other things) dumps of running jobs to disk which can be restarted from where they left off.
A look at the condor documentation might be order. The package is anon ftp'able.
-- -dave fetrow- INTERNET: fetrow@biostat.washington.edu FAX: 206-543-3286 BITNET: fetrow@uwalocke
NOTE: Sorry for the confusion. I am aware of the Condor package but the condor I mentioned in my posting was the name of the server on which the processes were running. Thanx - sud. ----------------------------------------------------------------------------
signal 2 is interupt, not 9 which is kill and can not be caught or ignored. what is parent process ? have you tried -TERM or -BUS ?
-heas -----------------------------------------------------------------------------
Find and kill the Parent Process.. if the parernt is INIT do a kill -HUP to init that should work...
########################################################################## + Richard J. Niziak + Systems Engineer + e:mail -> rickn@copley.com Copley Systems + land mail -> Copley Systems, Inc + 165 University Ave + Westwood, MA 02090 + voice mail -> (617)320-8300 x305 + ########################################################################## ----------------------------------------------------------------------------
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:28 CDT