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

Re: [Condor-users] Disk IO



Hey Matthew,

Thanks a lot for this. Now everything works perfectly.
Just wondering that condor is not "getting" the right size of the disks, do I need to configure it somewhere else?
I have mounted:
/dev/sdd /var/lib/condor/execute1
/dev/sdc /var/lib/condor/execute2
/dev/sdd /var/lib/condor/execute3
/dev/sde /var/lib/condor/execute4

All disk are 1 tera and are empty, but condor thinks they are full. Looks like it is getting the space from the EXECUTE and not the SLOT<N>_EXECUTE. The way I found to bypass it is adding:
request_disk = 0
requirements = $(TARGET.Disk>=0)

as requirements... Any idea?

On Fri, Jun 8, 2012 at 2:30 AM, Matthew Farrellee <matt@xxxxxxxxxx> wrote:
http://research.cs.wisc.edu/condor/manual/v7.6/3_3Configuration.html#15621

SLOT<N>_EXECUTE

Enjoy.

Best,


matt


On 06/07/2012 05:26 PM, Tiago Macarios wrote:
I dont really want to wait something for the jobs to start, just want
each job to be scheduled on a different disk and each disk having a 8
job limit. (8 jobs *4 disks= 32 the same as cores)

I dont think condor can have multiple EXECUTE folders, right? So I was
thinking about doing something like:
1 - somehow get the disk the job is suppose to run
2 - copy all files to the disk
3 - run it there
4 - copy back
5 - clean up

It just feels that it is something condor should do... disks are a
resource too, right? Something like a EXECUTE folder per disk and a
concurrency limit.

On Thu, Jun 7, 2012 at 5:02 PM, Diane Trout <diane@xxxxxxxxxxx
<mailto:diane@xxxxxxxxxxx>> wrote:

   The way we tried to deal with that was by writing a small daemon and
   helper thatinjected the fileservers load into condor.

   We then also modified the START conditional to make sure the
   fileservers load was reasonable.

   # this condor block tries to grab the load information:
   STARTD_CRON_JOBLIST = $(STARTD_CRON_JOBLIST) fileserver_load
   STARTD_CRON_FILESERVER_LOAD_MODE = Periodic
   STARTD_CRON_FILESERVER_LOAD_EXECUTABLE = /usr/bin/get_load
   STARTD_CRON_FILESERVER_LOAD_ARGS = "-options"
   STARTD_CRON_FILESERVER_LOAD_PERIOD=60s
   STARTD_CRON_FILESERVER_LOAD_LOAD=1

   # The start conditionals look like:
   CpuBusy = ((LoadAvg / TotalCpus) > 1.5) || (fileserver_load > 9.0)

   START=(CpuBusy == FALSE)

   After implementing the above solution I then learned about the
   concurency limits which might also work.

   http://research.cs.wisc.edu/condor/manual/v7.4/3_13Setting_Up.html#SECTION0041314000000000000000

   In some magical land condor would be able to what file servers a job
   is hitting and throttle just those jobs.

   On Thu, Jun 07, 2012 at 04:47:40PM -0400, Tiago Macarios wrote:
    > Hi All,
    >
    > I have a particular problem, if someone could help me I would
    > really appreciate.
    > I have a condor pool of approx 30 machines, this machines run
   jobs that are
    > disk and CPU intensive, but lately the jobs are more disk IO
   intensive than
    > CPU intensive. This is causing some machines to actually IDLE
   while the
    > disk is seeking. These machines can run 32 processes at a time, but
    > currently I have to enforce a limit of 8 per machine (heuristic
   value),
    > because of the disk.
    >
    > The machines already have 4 disks each. I was wondering if there
   is a way I
    > could create a concurrency limit for disks and make condor copy
   and run the
    > files on the different disks enforcing the 8 jobs per disk. The
   problem is
    > that how can I get a variable that tells me which disk I should
   copy things
    > to?
    >
    > Thanks,
    >
    > Mac.

    > _______________________________________________
    > Condor-users mailing list
    > To unsubscribe, send a message to
   condor-users-request@xxxxxxxxedu
   <mailto:condor-users-request@cs.wisc.edu> with a

    > subject: Unsubscribe
    > You can also unsubscribe by visiting
    > https://lists.cs.wisc.edu/mailman/listinfo/condor-users
    >
    > The archives can be found at:
    > https://lists.cs.wisc.edu/archive/condor-users/


   _______________________________________________
   Condor-users mailing list
   To unsubscribe, send a message to condor-users-request@xxxxxxxxedu
   <mailto:condor-users-request@cs.wisc.edu> with a

   subject: Unsubscribe
   You can also unsubscribe by visiting
   https://lists.cs.wisc.edu/mailman/listinfo/condor-users

   The archives can be found at:
   https://lists.cs.wisc.edu/archive/condor-users/




_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxedu with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/