On Sun, 06 Feb 2011 23:02:13 -0500, Dan Langille <dan AT langille DOT org> 
wrote:
> I recently changed how I do job copy.  I used to do them all at once. 
> and frequently.  Now I've moved to doing them in batches, once a day. 
> All my incr, all my diff, and all the fulls.
> 
>    http://www.freebsddiary.org/bacula-disk-to-tape-via-sql.php
> 
> Why?  Less tape churn.  Previously, the copy-all-from-pool approach was 
> mixing up different backup levels.  Thus, you might have an inc, diff, 
> inc, full, inc, diff... resulting in several tape changes.  By running 
> one copy job for just inc, then another copy job for diff, etc, the 
> tapes churn less.

I am doing something similar on my home setup. I have 3 Tape Drives (1
LTO-1 for Full, 1 DLTIIIXT for Incremental and 1 DLTIII for Differential).
Also I have 1 file pool each for Full, Incr, Diff with the tape drives as
"Next Pool". I have jobs running every day at noon to migrate stuff based
on time onto the tapes. The times are set in a way to make sure I always
have the latest Full (and Differential) and the necessary Incremental
backups on file storage for fast restore.

Next thing I want to try out is putting the File Pools on DragonFly BSD
with HAMMER filesystem. HAMMER supports Data De-Dup in batch job and also
in live-mode on the development branch.

Jan

-- 
professional: http://www.oscar-consult.de
private: http://neslonek.homeunix.org/drupal/

------------------------------------------------------------------------------
The modern datacenter depends on network connectivity to access resources
and provide services. The best practices for maximizing a physical server's
connectivity to a physical network are well understood - see how these
rules translate into the virtual world? 
http://p.sf.net/sfu/oracle-sfdevnlfb
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users