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

Re: [Condor-users] starter error

Michael Thomas wrote:
I'm setting up a new Condor 7.2.0 cluster on 64-bit Centos 5.2.  When I
submit a job to the local schedd, the job gets scheduled but fails to
run on the execution host.  The error in the starter log on the execute
host is:

2/6 16:56:02 warning: unable to chmod condor_exec.exe to ensure execute
bit is set: Operation not permitted
2/6 16:56:02 Job 6.0 set to execute immediately
2/6 16:56:02 Starting a VANILLA universe job with ID: 6.0
2/6 16:56:02 IWD: /wntmp/condor/compute-14-1/execute/dir_24649
2/6 16:56:02 Output file:
2/6 16:56:02 Error file:
2/6 16:56:02 Using wrapper /opt/condor/bin/condor_nfslite_job_wrapper.sh
to exec 30

...and the error that is returned to the user is:

/opt/condor/bin/condor_nfslite_job_wrapper.sh: line 14:
/wntmp/condor/compute-14-1/execute/dir_24649/condor_exec.exe: Permission
/opt/condor/bin/condor_nfslite_job_wrapper.sh: line 14: exec:
/wntmp/condor/compute-14-1/execute/dir_24649/condor_exec.exe: cannot
execute: Permission denied

Since the earlier chmod fails, my job wrapper script fails to 'exec' it
due to the missing execute permissions.  But I have no idea why the
initial chmod is failing.


You might have a problem if /wntmp is NFS with root_squash in effect.