Using CONDOR batch system for ATLAS jobs
Requirements
- NICADD cluster account
NICADD T3 cluster
Currently there are five nodes at the NICADD cluster to be used primarily for Atlas related analyses :
- pt3int0 (t3int0.nicadd.niu.edu) - the Interactive public node,
for code development , batch submission and data sample downloads - pt3wrk0-pt3wrk4 - T3 worker nodes (176 batch slots total)
Additionally nodes pncX and phpcX nodes can run Atlas batch jobs (420 more slots) over data from /xdata or /bdata disks.Each node has a local disk space (/disk) that should be used as a scratch for batch jobs.
CONDOR batch system
Please read the CONDOR user manual , specifically the "submitting a job" and "managing a job" chapters. The important commands are condor_status, condor_q, condor_submit, condor_hold, and condor_rm.
Here are the examples of job description file (Attach:atlasT3_CondorCmd.template) and shell script (Attach:atlasT3_jobtemplate.sh) that uses the ATLAS environment.To start, download copy both files into a test area:
mkdir /xdata/$USER/condor_test cd /xdata/$USER/condor_test cp atlasT3_CondorCmd.template atlasT3_Condor.cmd
//edit atlasT3_Condor.cmd, replace "user" with actual username
//edit atlasT3_jobtemplate.sh, customize for the real job
//run atlasT3_jobtemplate.sh locally, if satisfied submit job
condor_submit atlasT3_Condor.cmd