Mar 302014
 

Last night I started to catch up on copying jobs from disk to tape. There were about 420 incremental jobs waiting. I started my CopyToTape-Inc job, and it queued up 417 jobs:

30-Mar 02:17 bacula-dir JobId 167491: The following 417 JobIds were chosen to be copied: 165711,165712,165724,165726,165728,165730,165731,165734,165735,165740,165741,165743,165744,165746,165749,165753,165755,165758,165761,165762,165727,
165729,165732,165736,165742,165747,165754,165756,165759,165733,165751,165757,165760,165752,165739,165775,165776,165830,165831,165832,165834,165835,
165838,165839,165844,165845,165847,165848,165850,165853,165857,165859,165862,165865,165866,165836,165840,165846,165851,165854,165858,165863,165837,
...
...
...
167393,167368,167373,167379,167382,167386,167391,167374,167383,167389,167392,167384,167371

The next morning, all jobs had completed, but there were still jobs left to be copied over (as determined by my magical Bacula portal).

I ran the job again. Let them finish. Still more jobs left to job.

I talked on IRC and found 100 seemed to be the maximum number of jobs that one could queue up at one time. A mailing list post and the source code reference it refers to confirms this.

Starting with Bacula 7.0, this limit is controlled by a configuration setting, Maximum Spawn Jobs. The migrate.c functionality is now split between mac.c (for migrate and copy) and mac_sql.c. You can see the configuration setting in the code at this link.

I will be working on creating a new Bacula port for FreeBSD as soon as I can. In them meantime, I’ll just keep re-running my copy job. I’m looking forward to running Bacula 7.

Website Pin Facebook Twitter Myspace Friendfeed Technorati del.icio.us Digg Google StumbleUpon Premium Responsive