[GE users] job_load_threshold under 6.0u1

Sean Dilda agrajag at dragaera.net
Wed Oct 6 21:32:56 BST 2004

I'm trying to use job_load_threshold under 6.0u1 and it doesn't seem to
be working.  Before I file a bug on it, I'd like to make sure this isn't
a known issue, and make sure I'm not doing something wrong.

In my sched_conf I have:
job_load_adjustments              np_load_avg=1.0
load_adjustment_decay_time        0:7:30

In the queue I'm using, I have:
load_thresholds       np_load_avg=0.9

My test job is a script that merely sleeps for 500 seconds, then exits.
As I understand the docs, this means that as soon as the job starts
running, the host in question will hit alarm state, but only for a
matter of seconds.  I submit the job, then repeatedly start running
'qhost -q'.  I see the job enter the queue, but the queue never hits the
alarm state.  I also try using 'qhost -F np_load_avg', however the
np_load_avg for the machine in question never goes above 0.0000000

I know that the load_thresholds are working correctly, but the load
adjustment doesn't seem to.  Has anyone else seen this?  Is there
something I'm doing wrong?



To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net

More information about the gridengine-users mailing list