[torqueusers] Re: kill_delay

Roy Dragseth Roy.Dragseth at cc.uit.no
Tue Feb 27 02:50:42 MST 2007


After some tinkering with the code I've come to the conclusion that the kill 
loop makes a lot of sense for parallel jobs, as you want to give an mpi 
launcher the time to clean up before it is killed with an untrappable signal.  
The loop is only executed on a SIGKILL.  The annoying delay should be fixed 
by doing a fork.

Anyway, the kill_delay parameter is only influencing the delay between SIGTERM 
and SIGKILL when you do a qdel, it does not have any influence on the delay 
when a job hits the walltimelimit.  This is stated in the man-page:


kill_delay
                 The  amount  of the time delay between the sending of SIGTERM 
and SIGKILL when a qdel command is issued against a running
                 job.  This is overriden by the execution queue attribute of 
the same name. Format: integer seconds; default value: 2 sec-
                 onds.

As far as I can test kill_delay works correctly on a qdel.

r.

-- 

  The Computer Center, University of Tromsø, N-9037 TROMSØ Norway.
	      phone:+47 77 64 41 07, fax:+47 77 64 41 00
     Roy Dragseth, High Performance Computing System Administrator
	 Direct call: +47 77 64 62 56. email: royd at cc.uit.no


More information about the torqueusers mailing list