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

Re: [HTCondor-users] Getting batch-queue wait time of Condor pool and the number of free nodes



On Wed, Sep 11, 2013 at 11:54 PM, Harshad Prajapati
<harshad.b.prajapati@xxxxxxxxx> wrote:

> How is it possible to get following through program and remotely?
>
> Can we get batch-queue wait time of any Condor pool?

It depends on what you mean by wait time. If you want to know the time
a job will have to wait until it runs, that's not how HTCondor works
(unlike, say, PBS). You can see the time that jobs have been in the
queue using the appropriate condor_q variation. For example, to see
the time (Unix epoch time) that idle jobs were submitted the local
queue:
  condor_q -const 'JobStatus==1'  -format '%i\n' QDate

To another queue in your pool:
  condor_q -name otherschedd -const 'JobStatus==1'  -format '%i\n' QDate

To another queue in another pool:
 condor_q -name otherschedd -pool htcondor.example.com -const
'JobStatus==1'  -format '%i\n' QDate

> Can we get the number of free nodes of any Condor pool?
>
The condor_status command can provide that. You can get a listing of
all the nodes in a pool with no arguments. You can get the summary
counts for all nodes with `condor_status -total`. Or you can get just
the unclaimed nodes with `condor_status -const 'State=="Unclaimed"'`
(you can use -total or not, depending on what sort of information you
want). Again, for a remote pool, use the -pool argument.

You may find the following man pages very helpful:
http://research.cs.wisc.edu/htcondor/manual/latest/condor_q.html
http://research.cs.wisc.edu/htcondor/manual/latest/condor_status.html

> Is above monitoring inbuilt in Condor? or Is there any resource monitoring
> infrastructure which monitors Condor pool for these two information?
>
CycleComputing has a tool called CycleServer
(http://cyclecomputing.com/cycleserver/overview) that provides
monitoring and management tools for HTCondor. There is also a tool
called Cumin (https://fedorahosted.org/grid/wiki/Cumin). Some people
continue to use CondorView
(http://research.cs.wisc.edu/htcondor/manual/latest/9_4HTCondorView_Client.html)
to monitor their pools.


Thanks,
BC

-- 
Ben Cotton
Senior Support Engineer
Cycle Computing, LLC
The Leader in Utility Supercomputing Software