| 28 |
|
-kill [range] -- kill submitted jobs. |
| 29 |
|
-resubmit [range] -- resubmit killed/aborted/retrieved jobs. |
| 30 |
|
-forceResubmit [range] -- resubmit jobs regardless to their status. |
| 31 |
< |
-copyData [range [dest_se or dest_endpoint]] -- copy locally (in crab_working_dir/res dir) or on a remote SE your produced output, |
| 32 |
< |
already stored on remote SE. |
| 31 |
> |
-copyData [range [dest_se or dest_endpoint]] -- copy locally (in crab_working_dir/res dir) or on a remote SE your produced output, |
| 32 |
> |
already stored on remote SE. |
| 33 |
|
-renewCredential -- renew credential on the server. |
| 34 |
|
-clean -- gracefully cleanup the directory of a task. |
| 35 |
|
-match|-testJdl [range] -- check if resources exist which are compatible with jdl. |
| 256 |
|
|
| 257 |
|
Select the ARC scheduler: |
| 258 |
|
scheduler = arc |
| 259 |
< |
|
| 259 |
> |
|
| 260 |
|
=item B<requirements>, B<additional_jdl_parameters:> |
| 261 |
|
|
| 262 |
|
Use xrsl code instead of jdl for these parameters. |
| 263 |
|
|
| 264 |
< |
=item B<max_cpu_time>, B<max_wall_clock_time:> |
| 264 |
> |
=item B<max_cpu_time>, B<max_wall_clock_time:> |
| 265 |
|
|
| 266 |
|
For parameters max_cpu_time and max_wall_clock_time, you can use |
| 267 |
|
units, e.g. "72 hours" or "3 days", just like with the xrsl attributes |
| 529 |
|
|
| 530 |
|
See also I<increment_seeds>. Seeds not listed in I<increment_seeds> or I<preserve_seeds> are randomly set for each job. |
| 531 |
|
|
| 532 |
< |
=item B<first_run> |
| 532 |
> |
=item B<first_lumi> |
| 533 |
|
|
| 534 |
< |
Relevant only for Monte Carlo production for which it defaults to 1. The first job will generate events with this run number, subsequent jobs will |
| 535 |
< |
increment the run number. Failing to set this number means CMSSW will not be able to read multiple such files as they |
| 536 |
< |
will all have the same run and event numbers. This check in CMSSW can be bypassed by setting |
| 534 |
> |
Relevant only for Monte Carlo production for which it defaults to 1. The first job will generate events with this lumi block number, subsequent jobs will |
| 535 |
> |
increment the lumi block number. Setting this number to 0 (not recommend) means CMSSW will not be able to read multiple such files as they |
| 536 |
> |
will all have the same run, lumi and event numbers. This check in CMSSW can be bypassed by setting |
| 537 |
|
I<process.source.duplicateCheckMode = cms.untracked.string('noDuplicateCheck')> in the input source, should you need to |
| 538 |
< |
read files produced without setting first_run. |
| 538 |
> |
read files produced without setting first_run (in old versions of CRAB) or first_lumi. |
| 539 |
|
|
| 540 |
|
=item B<generator> |
| 541 |
|
|
