Ciao David,
> what should I do with the 100k files in
> /opt/edg/var/gatekeeper/jobs accumulated until now?
> Will they be reprocessed somehow, or are all the accounting information lost?
They will not be processed automatically. Worse, there is a cron job
that deletes any such file that is older than a week:
/etc/cron.d/cleanup-job-records
Each file misses one essential ingredient: the batch system job ID.
For each grid job ID (included if the job was submitted via a _WMS_)
you can determine the batch system job ID from /var/log/messages*
and add it to the corresponding line; then add each such line to the
correct grid-jobmap_201204* file and let DGAS or APEL reprocess the
affected files.
A lot of work - maybe not really worth the effort...
|