Re: [Condor-users] Quill database stays empty

[Ameet Kini <akini@xxxxxxxxxxx>]

> I'm getting a stong sense of dumb crawling up my spine.  I've been
> thinking of Quill as a centralized thing -- a database copy of jobs as
> they are matched by the central manager.  I have a quill daemon running
> only on that machine.  Is it supposed to be run on every submit node?

Quill runs alongside the schedd. It sniffs the job queue log file written
by the schedd.  So yeah, it needs to be on the same machine as the schedd.

>
> Nothing is being logged to the schedd log on that one node because none
> of the submissions are coming from that node.  The file is larger than
> 47 bytes, but it hasn't changed in months.
>
> - dave
>
>
> Ameet Kini wrote:
> >
> >>initial creation of the database.  The Quill daemon is running on the
> >>same machine as the negotiator and collector and writing to a postgres
> >
> >
> > And the schedd along with the negotiator and collector, I assume?
> >
> > Seems like its able to find a job queue log file. And seems like the
> > job queue log file is practically empty (47 bytes usually corresponds
> > to the first two header lines).  But you mentioned that the schedd *is
> > writing* a bunch of records to the same job queue log as the one that is
> > sniffed by Quill - can you double check that? If the schedd isn't doing
> > anything to that job queue log, then Quill will detect no activity.
> >
> > Ameet
> >
> > On Tue, 13 Dec 2005, David A. Kotz wrote:
> >
> >
> >>These are the only sort of messages that have been logged since the
> >>initial creation of the database.  The Quill daemon is running on the
> >>same machine as the negotiator and collector and writing to a postgres
> >>server on a separate machine.
> >>
> >>
> >>
> >>12/13 09:41:30 ******** Start of Probing Job Queue Log File ********
> >>12/13 09:41:30 === Stored Last Probing Information ===
> >>12/13 09:41:30 fsize: 47                mtime: 1132675007
> >>12/13 09:41:30 === Current Probing Information ===
> >>12/13 09:41:30 fsize: 47                mtime: 1134168248
> >>12/13 09:41:30 POLLING RESULT: NO CHANGE
> >>12/13 09:41:30 ********* End of Probing Job Queue Log File *********
> >>12/13 09:41:30 ++++++++ Sending schedd ad to collector ++++++++
> >>12/13 09:41:30 ++++++++ Sent schedd ad to collector ++++++++
> >>12/13 09:41:40 ******** Start of Probing Job Queue Log File ********
> >>12/13 09:41:40 === Stored Last Probing Information ===
> >>12/13 09:41:40 fsize: 47                mtime: 1132675007
> >>12/13 09:41:40 === Current Probing Information ===
> >>12/13 09:41:40 fsize: 47                mtime: 1134168248
> >>12/13 09:41:40 POLLING RESULT: NO CHANGE
> >>12/13 09:41:40 ********* End of Probing Job Queue Log File *********
> >>12/13 09:41:40 ++++++++ Sending schedd ad to collector ++++++++
> >>12/13 09:41:40 ++++++++ Sent schedd ad to collector ++++++++
> >>12/13 09:41:50 ******** Start of Probing Job Queue Log File ********
> >>12/13 09:41:50 === Stored Last Probing Information ===
> >>12/13 09:41:50 fsize: 47                mtime: 1132675007
> >>12/13 09:41:50 === Current Probing Information ===
> >>12/13 09:41:50 fsize: 47                mtime: 1134168248
> >>12/13 09:41:50 POLLING RESULT: NO CHANGE
> >>12/13 09:41:50 ********* End of Probing Job Queue Log File *********
> >>12/13 09:41:50 ++++++++ Sending schedd ad to collector ++++++++
> >>12/13 09:41:50 ++++++++ Sent schedd ad to collector ++++++++
> >>
> >>
> >>
> >>
> >>Ameet Kini wrote:
> >>
> >>>Can you cut & paste any relevant parts of your QuillLog file?  Otherwise,
> >>>its tough to tell what exactly is going on.
> >>>
> >>>--Ameet
> >>>
> >>>On Mon, 12 Dec 2005, David A. Kotz wrote:
> >>>
> >>>
> >>>
> >>>>I've recently attempted to set up a test pool to use Quill.  It works up
> >>>>to the point of creating the database structure, but it never writes any
> >>>>data to the database.  When I connect to the database and query it using
> >>>>psql, the tables are empty.  The tables are created, so the database and
> >>>>the quillwriter user seem to be configured correctly.  There are plenty
> >>>>of jobs running through the queue, so there should be data.  I'd
> >>>>appreciate any suggestions as to where I might have gone wrong.
> >>>>
> >>>>- dave
> >>>>_______________________________________________
> >>>>Condor-users mailing list
> >>>>Condor-users@xxxxxxxxxxx
> >>>>https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> >>>>
> >>>
> >>>
> >>>Ameet
> >>>_______________________________________________
> >>>Condor-users mailing list
> >>>Condor-users@xxxxxxxxxxx
> >>>https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> >>
> >>_______________________________________________
> >>Condor-users mailing list
> >>Condor-users@xxxxxxxxxxx
> >>https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> >>
> >
> >
> > Ameet
> > _______________________________________________
> > Condor-users mailing list
> > Condor-users@xxxxxxxxxxx
> > https://lists.cs.wisc.edu/mailman/listinfo/condor-users
>
> _______________________________________________
> Condor-users mailing list
> Condor-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
>

Ameet