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

Re: [Condor-users] condor_rm not transferring the files back to submitter




You could use condor_vacate_job to abort the job. The output files will be transferred back to the job's spool directory. In versions of Condor from 7.5.5 and more recent, a job with id X.Y has its spool directory in $(SPOOL)/<X mod 10000>/<Y mod 10000>/clusterX.procY.subproc0, where the job id is X.Y. After you are done examining the spooled files, you can remove the job.

Another way to examine job output for a running job is to use condor_ssh_to_job (in unix, not windows). This can be done to interactively look at files or to fetch them with scp or rsync.

--Dan

On 7/14/11 7:18 AM, sassyn@xxxxxxxxx wrote:
I see, thank u. Is there any other way to do that? My jobs write a lot of files, and sometimes i need to kill the jobs, since i see there are errors in the main log file (i use the stream option). But then i want to check the other log files using some tools automatic tools i develope, but then the files are not exits. Surely i can wait the job to complete, but this is just a waste of time and resources. I guess there should be a work around for this. Do u have an idea? Thank you. Sassy
-----Original Message-----
From: Matthew Farrellee
Sent:  14/07/2011, 14:59
To: Condor-Users Mail List
Subject: Re: [Condor-users] condor_rm not transferring the files back to	submitter


On 07/14/2011 07:12 AM, Sassy Natan wrote:
Dear Group,

I have submitted the following job into the queue of condor"

/executable = /usr/local/bin/matlab_divx_pro<http://chip_sim.pl/>/
/should_transfer_files = YES
when_to_transfer_output = ON_EXIT_OR_EVICT
universe = vanilla
getenv = True
notification = Error
run_as_owner = True
load_profile = True
initialdir =/usr/local/bin/
concurrency_limits = matlab
transfer_output_files = output
stream_error = true
stream_output = true
priority = 0
args =    -no_sva   -report Report_2011-07-14_13-06-14_c_//example
Error = log.err
Output = log.out
Log = log.log
Queue/

However, when removing the job from the q using condor_rm I want to get
the log files, and output files back to submit machine.

In the condor documents I see that /"when_to_transfer_output =
ON_EXIT_OR_EVICT"/ it should//"always do the transfer, whether the job
completes on its own, is preempted by another job, vacates the machine,
or is killed." (see here
http://www.cs.wisc.edu/condor/manual/v7.6/2_5Submitting_Job.html#SECTION00354200000000000000)

But when doing condor_rm to the specific job I don't get the files back.

Can someone help or drop a comment?


Thanks
Sassy
"EVICT" in "ON_EXIT_OR_EVICT" refers to when policy on the slot machine
initiates the eviction, not the user.

Best,


matt
_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx 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@xxxxxxxxxxx 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/