On Wed, Apr 22, 2009 at 02:01:06PM -0500, Jeremy Parrish wrote: > >Works fine for me without doing a pool per USB HDD. If bacula can't > >find the old volume, it makes a new one with the same name, which > >seems a little odd, but works OK.
> This is what I am concerned about. When it creates a new volume with > the same name as a volume that exists on another drive, things get > confusing because 1) The "old" volume becomes wasted space on another > drive, 2) Tracking down which drive you need based on volume name > becomes nearly impossible, and 3) Bacula can become confused if it > thinks it is looking at the the "re-created" volume but is actually > looking at the "orphaned" volume on another drive. > > Do you have a solution for these issues, or have they not arisen for > you? (1) "wasted space" on the old drive is not a problem for me. When the drive is not active, it's not like I need that space for other things. When the drive comes back for reuse, the whole drive needs to be wiped anyway, or bacula can be confused. (2) Tracking down the drive based on the volume name is a concern for me, although it hasn't yet been a problem in practice [yet?] (3) confusion between stale old volume names and current volume names can exist in two scenarios -- when reusing an old drive, or when doing a restore. I have workarounds for both. When reusing an old drive, do an mkfs first -- you're anyway preparing to reuse it. This eliminates confusion. When restoring from an old drive with some stale volumes and some current, you are probably specifying "before a specified time". bacula will think the stales are *new*, so it won't be confused -- although you might be. I don't much like it, but it seems to work OK. Migration/copy might be a better solution for the underlying design goal. - Morty ------------------------------------------------------------------------------ Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users