[HTCondor-users] One job per machine

[Michael Thomas <wart@xxxxxxxxxxx>]

I've got a workflow that is very network intensive (verifying file 
checksums from a distributed ceph filesystem).  Two jobs running on the 
same machine in the cluster can easily saturate a cluster node's 1Gbps 
network connection.

Is there some constraint I can use in my submit file to say "I only want 
one instance of this job to run on any startd" in order to avoid 
saturating an individual node's network connection?

executable = /home/michael.thomas/condor/cephcheck/md5check.sh
args = $(md5file)
universe = vanilla
accounting_group=llo.test
request_disk = 1000
request_memory = 1000
log = /home/michael.thomas/condor/cephcheck/test.log
error = /home/michael.thomas/condor/cephcheck/test.err.$(Process)
output = /home/michael.thomas/condor/cephcheck/test.out.$(Process)
max_materialize = 10
queue md5file from md5_list.txt

--Mike