[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Master scheduling Intervals for debugging

On 8/20/07, Mills, Ashley, VF-Group <Ashley.Mills@xxxxxxxxxxxx> wrote:
> I am having a daymare (as opposed to a nightmare) getting my scripts to
> execute without errors. I would like to debug what is going on, but
> there is a huge delay between submitting a job, and when it starts
> running (perhaps 10 minutes).


Here is what I do.  I run the following in one window:

'watch condor_q'

this shows me the queue and running status of jobs

in another window I do this:

condor_submit myjob

and then, in another window, I do:


This forces the condor daemons to reschedule more immediately.
Sometimes it works,
sometimes it does not.

If you are using DAGman, expect to wait about 60 seconds before the
first job in the DAG
executes no matter how many times you run condor_reschedule.

I have complained to the condor developers multiple times about the
fact that the granularity
of the scheduling daemon is too long and why it takes so long is
completely opaque.  No matter how you configure the timeouts,  there
are somethings that just don't happen quickly enough.