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

Re: [HTCondor-users] Permission denied (errno 13) only on submitter machines



The default value for file transfer in a submit file is

    should_transfer_files = IF_NEEDED

This means that the files will be transferred if the HTCondor determines that the STARTD and the SCHEDD don't have a shared file system.  (or don't have the same one). 

When the jobs execute on the submit node, then HTCondor sees that they DO share a file system, and so it tries to run the job from the submit directory rather than transferring the files into the execute directory.  This is more efficient,  but the default permissions for c:\users does not permit this unless the jobs are also set to run_as_owner, so the jobs will fail to start. 

You can force file transfer to happen even when there is a common file system between submit and execute by adding this line to your submit file.

  should_transfer_files = YES

-tj

-----Original Message-----
From: HTCondor-users [mailto:htcondor-users-bounces@xxxxxxxxxxx] On Behalf Of Julio Augusto Freyre-Gonzalez
Sent: Sunday, May 13, 2018 12:22 PM
To: htcondor-users@xxxxxxxxxxx
Subject: [HTCondor-users] Permission denied (errno 13) only on submitter machines

I'm managed to configure a condor installation in a pool of four Windows
multicore workstations totalizing 62 cores. The four workstations can
submit works. When I send a cluster of 62 jobs (in the vanilla
universe), each is executed in each machine without any problems except
for the jobs assigned to the submitter machine. These jobs stay on hold.

condor_q -analyze reports "Hold reason: Error from slot1@ADAMANTIUM:
Failed to open 'C:\Users\mtello\Desktop\test\testout50.txt' as standard
output: Permission denied (errno 13)"

These are also the only errors in StarterLog.* files:
Failed to open 'C:\Users\mtello\Desktop\test\testout50.txt' as standard
output: Permission denied (errno 13)
Failed to open 'C:\Users\mtello\Desktop\test\error50.log' as standard
error: Permission denied (errno 13)

But again, these errors only occurs in the submitter machine (regardless
what workstation is used to submit). All the jobs assigned to other
machines runs without any problems. If I use the local universe, the
jobs also runs if any problems.

If I move the directory containing my submit files out of the user
account into a world-modifiable directory and I submit the jobs all of
them runs without any problems.

Additional info: I have set UID_DOMAIN to 'ccg.unam.mx' and all the
machines are in the same domain. But a condor_status -master reports not
fully qualified hostnames:

ADAMANTIUM   8.6.11.440910     8     7.9 GB    0+01:00:04
BENZAITEN    8.6.11.440910     8     7.9 GB    0+01:00:06
XANADU       8.6.11.440910    32    31.9 GB    0+10:50:08
ZARATHUSTRA  8.6.11.440910    24    15.9 GB    0+01:00:05

Thanks for your help.
Best,
Julio
-- 
Dr. Julio Augusto Freyre-Gonzalez (PhD)
Associate Professor
Head of the Regulatory Systems Biology Research Group
http://freyrelab.org
Evolutionary Genomics Program
Center for Genomic Sciences
National Autonomous University of Mexico (UNAM)
Av. Universidad s/n, Col. Chamilpa, 62210
Cuernavaca, Morelos. Mexico

"Ser joven y no ser revolucionario es una contradiccion hasta biologica"
- Salvador Allende Gossens (1908-1973)

"If we knew what it was we were doing, it would not be called research,
would it?"
- Albert Einstein (1879-1955)

"There is one thing even more vital to science than intelligent methods;
and that is, the sincere desire to find out the truth, whatever it may be"
- Charles Sanders Pierce (1839-1914)

"The great tragedy of science -- the slaying of a beautiful hypothesis
by an ugly fact"
- Thomas Henry Huxley (1825-1895)
_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

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