mvicenzi / pmt-calibration Goto Github PK

Older data is stored on tape and requires prestaging before submitting jobs. Submission scripts should be updated to handle these cases. If not prestaged, files on tape are prestaged individually when the job tries to access them. This leads to low efficiency for jobs (long wait times).

Things to do:

• Move away from input file lists, using samweb dataset definitions. (or keep both options?)
• Check if files are on disk or not before submitting
• Add prestaging before submitting (if necessary)