Hi, We currently have an instance of bacula running with ~200 clients running on a weekly schedule (one backup per day, 2 full per week and incremental between the full runs) that totals ~6TB compressed backup data. Due to the relatively large amounts of data and the lack of speed of certain clients we have chosen to enable multiple concurrent jobs (4 per pool). Doing a full run for all clients can take well over 48 hours without multiple concurrent jobs enabled.
All backups are to disk, so no tapes are involved at all. We currently have all pools 'volume use duration' set to 23 hours to avoid creating to large volume files. This gives us individual volume sizes of ~250GB for our largest pool. Anyway, the problem we've been having is that after a volume rotates we receive errors similar to: 01-Mar 01:16 birch-dir: Start Backup JobId 15091, Job=[xxx].2007-03-01_01.08.44 01-Mar 01:16 birch-dir: Recycled volume "Vol-windows-940001" 01-Mar 01:16 birch-sd: [xxx].2007-03-01_01.08.44 Fatal error: acquire.c:355 Wanted Volume "Vol-windows-940001", but device "FileStorage01-Windows-94" (/backup/data/01/bacula-data/fs01-windows-94/) is busy writing on "Vol-windows-940013" . This happens for all jobs that try to run while a job is still keeping the old (pre-rotated) volume open. Once the 'old' job finishes the device is able to mount the new volume and subsequent jobs run as intended. Any suggestions for how to fix this? Either with or without concurrent jobs? One solution would obviously be to disable concurrent jobs, but that would result in a full run taking far to long time. Another possible solution would be to raise the volume use duration to avoid volume cycling for longer periods, but that would really only make problem less frequent and it would also result in even larger volumes. Ideally I'd even like to lower the volume use duration by quite a bit as bacula seems to read through the volume a backup is stored on sequentially when doing a restore, and this can take quite a while for large volumes. But lowering it would result in more volume rotations which would result in more errors, so I can't really do that as things currently stand. Any advice/suggestions appreciated. Regards, -- Simon Ekstrand ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
