Setup a project.py job - twongjirad/LArLiteSoftCookBook GitHub Wiki

The answer

An example is found in examples/project.py

Same idea, but a MC example is here

Running It

  • First make sure you start up a proxy
  • to submit

project.py --xml [myproject].xml --stage [stage1] --submit

  • Check jobs

project.py --xml [myproject].xml --stage [stage1] --check

  • Resubmit failed jobs

project.py --xml [myproject].xml --stage [stage1] --makeup

  • Nuke the stage's folders

project.py --xml [myproject].xml --stage [stage1] --clean

annotated

Some things to note

  • datadir should go into dcache (/pnfs/uboone/scratch/users/[me])
  • workdir goes into bluearc (/uboone/data/users/tmw)
  • You need to make a tarball for your copy of the code and specify it's location. After setting up a larsoft/uboonecode environment, go into the base directory of the code and run

make_tar_uboone.sh [name of tarball file]

e.g. If you are running uboonecode v04_27_01 then you might type

make_tar_uboone.sh localProducts_larsoft_v04_27_01_e9_prof science_magic.tar

  • After making tarball, be sure to specify your tarball in the xml file by adding/changing the following * line. e.g. If you were me (Jarrett in this case) and had just tar'd as above and made science_magic.tar you'd do

<larsoft> <local>/uboone/app/users/moon/OpticalStudies/v04_27_01/science_magic.tgz </local> </larsoft>

  • Monitor jobs and batch server here
  • to check jobs in queue:

jobsub_q --user [userid]

  • To get logs

jobsub_fetchlogs --jobid=[jobid]

References

http://microboone-docdb.fnal.gov:8080/cgi-bin/RetrieveFile?docid=4020&filename=IIT_20150106.pdf&version=2

⚠️ **GitHub.com Fallback** ⚠️