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

Re: [HTCondor-users] looking for a condor_q job control gui or web app

On 2/4/2016 12:39 PM, Mary Romelfanger wrote:

Hello Everyone,

I am hoping to not need to reinvent the wheel for this problem.

We are in finally in the process of moving our HTCondor based system
into our realtime operations pipeline.  This means that it is being
handed over to operators that have various levels of linux skills.  In
the previous data processing pipeline this varying level of command line
skills has worked because all operator tasks have always been gui based.
    We have developed a new data processing pipeline with HTCondor at
its heart and with a web app and database structure that allows the
operators to monitor the datasets as they move through the many steps in
our system.

Hi Mary,

The above is great news, congrats!  More below...

But the one piece that we are missing is the ability to monitor the
condor_q, issue hold, release, and rm commands for specific jobs, and
modify the job priority setting  via a gui or web application rather
than the command line.

Log parsing would be an nice side benefit, but the job control options
are the critical requirement.  Do any of you have any suggestions?
  Did HTCondorView have this ability, or is it just a log parser?  But
it appears to be going away, anyway?

HTCondorView is not going away per se, but it is being rewritten in this developer series to be more flexible and not rely upon java applets (it is being redone via javascript + ajax). However, HTCondorView is not about submitting/managing the job queue but instead is all about displaying utilization and user usage information - think of HTCondorView as displaying similar sort of info that you could today with condor_gangliad + Ganglia, but without the need to install/configure anything.

As for a GUI that allows manipulation of the job queue, in addition to the solution Michael already mentioned from Red Hat, you may want to consider CycleServer from Cycle Computing. See
Looking at the Manual for CycleServer, it certainly appears to be able to manipulate jobs in the queue, see


I imagine there are users of both CycleServer and/or Cumin on this email list that could provide additional input, not to mention both Cycle and Red Hat folks often post here.

best regards,

Todd Tannenbaum <tannenba@xxxxxxxxxxx> University of Wisconsin-Madison
Center for High Throughput Computing   Department of Computer Sciences
HTCondor Technical Lead                1210 W. Dayton St. Rm #4257
Phone: (608) 263-7132                  Madison, WI 53706-1685