Configure → Job Batch Script collection

The job viewer tab in Open XDMoD is able to show the job batch script along with the other job information. The job batch scripts are stored in the database using the ingest_jobscripts.py process. The ingest_jobscripts.py tool is packaged with the Job Summarization software.

The ingestion of job batch scripts is an optional feature. It is not required for Open XDMoD to display job performance information.

The mechanism for extracting the job batch script is resource manager specific. Consult the documentation for your resource manager software for information about how to enable logging of the job batch script.

Source data schema

In order to ingest the job batch script data into Open XDMoD the job batch scripts must be stored in files. There must be one file per job and the job batch scripts for a given day should be stored in a datestamp named directory:

[BASEPATH]/YYYYMMDD/JOBID.savescript

where YYYY is the four digit year, MM two digit month, DD two digit day and JOBID is the identifier for the job from the resource manager. The date can refer to either the submit day, start day or end day of the job but must be the same for all jobs for a resource. The configuration settings for the path name and the datestamp meaning are described in the configuration guide.

The files are stored in datestamped directories because:

  1. Storing the files by date limits the number of files per directory.
  2. The datestamp is also used along with the job identifier to locate the accounting record for the job. The datestamp is used because job identifiers provided by a resource manager are typically not globally unique.
  3. The datestamp is used to limit the number of files to scan each time the ingest_jobscripts.py process runs.