i installed condor version 8.1.4 both on a linux workstation and a macbookÂ
i configured condor to allow jobs to be submitted from the macbook on the linux workstation and that works ok
however i would like to run parallel jobs using the parallel universe
i configured both installations according to the manual
i can query the status to see if the DedicatedScheduler is used and find on the linux workstation
$condor_status -long |grep Ded
and on the macbook as well
my jobs however remain in idle state and on the macbook side in the Schedlog i have these lines
04/27/14 18:48:32 (pid:27313) Sent ad to central manager for thomasge@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
04/27/14 18:48:32 (pid:27313) Sent ad to 1 collectors for thomasge@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
04/27/14 18:48:32 (pid:27313) Inserting new attribute Scheduler into non-active cluster cid=40 acid=-1
04/27/14 18:48:32 (pid:27313) Found 0 potential dedicated resources in 0 seconds
04/27/14 18:51:37 (pid:27313) Number of Active Workers 0
and on the linux box i get in CollectorLog
04/27/14 19:48:42 Got QUERY_STARTD_ADS
04/27/14 19:48:42 Number of Active Workers 0
04/27/14 19:48:42 (Sending 0 ads in response to query)
04/27/14 19:48:42 Query info: matched=0; skipped=4; query_time=0.001781; send_time=0.000156; type=Machine; requirements={( ( DedicatedScheduler == "DedicatedScheduler@xxxxxxxxxxxxxxxxxxxxxxxxxxxx" ) )}; peer=<
192.168.0.106:65216>; projection={}
it seems as if the macbook is looking for aÂDedicatedScheduler that is running under its own hostname.
putting a requirement in the submit file where i set the TARGETÂDedicatedScheduler to the one running on the linux workstation gives the same results
the output from condor_q -better-analyze indicates that all machines are ready to run the job.
best
Thomas