[GE users] 'dr' error state not clearing

Raymond Chan raychan at ucdavis.edu
Wed Jan 25 21:41:51 GMT 2006


Hi all,

 

I ran a parallel PVM job in SGE6 and did a qdel to abort it like usual.
Sometimes, after executing a qdel, the job doesn't disappear yet.  It
changes into a "dr" state which indicates it is registered for deletion.
After a little time, the job clears.  However, this time, the "dr' state is
sticking, and I don't know how to get rid of it.  I thought about trying to
delete the cluster queue the job was in, but it won't let me because a job
is "still running."  I can do qdel on the job as much as I can, but still
stuck in "dr".  I tried rebooting my head node and all cluster nodes as a
last resort, but the queue state stays as it is supposed to, and the "dr"
remains.

 

I'm a newbie, and hope it's something very simple to force a job delete some
how.  

 

Thanks,

Ray




More information about the gridengine-users mailing list