Hello,
Am 06.09.2005 schrieb Arno Lehmann:
> Hi,
>
> Masopust Christian wrote:
>
> >Hello all,
> >
> >my configuration is as follows:
> >
> > - Director and Storage at FedoraCore 3 system, Database Mysql 4.1.12,
> > Backup to Disk (2 RAIDs with 2.6TB)
> > - Clients are Linux, SUNs, Windows
> >
> >Bacula run fine till last weekend. Since Sunday every job fails with
> >messages
> >that entries to database ("INSERT INTO FILE (......") could not be made
> >because
> >of "Duplicate entries"....
>
> So the catalog is broken. Bad, but you should be able to correct it...
>
> >First i thought that i caused this problem myself as i run a "mysqldump" of
> >the whole bacual-db last week and after that i reloaded id (maybe dumb,
> >but i
> >heard that this could decrease size of db a little as mine is above 8GB).
>
> I don't know if this is necessary or useful, but that might have caused
> the problems.
It is sometimes necessary, that the database should be compacted some
day. I do it sometimes - but I realized, that I only get a few megabytes
in free space after re-importing the mysqldump-ed database then.
> >But i couldn't imagine that this causes my problems now because backup
> >runs fine
> >after the dump-reload for 2 days...
> >
> >so, any idea what happening here?
>
> Have you tried the dbcheck program to repair the database?
> I suggest to make sure that bacula is not running when you modify the
> database.
>
> Be aware that the database check ca take very long.
>
> You could also start with a new, completely empty catalog, and reinsert
> your existing volumes using bscan.
>
>
> >thanks a lot,
> >chris
> >
> >P.S.: btw, would it be better to have more catalogs when size of 1 catalog
> >is
> > as big is mine is?
> >
>
> I don't know, but I'm quite sure that using more than one catalog can
> result in some difficulties with you bacula setup.
The only thing what can be done to reduce the size of the database, is
using the "right" values for "Volume Retention", "File Retention" and
"Job Retention". For example, if you make a weekly change backup with 4
pools, why should be e.g. "Volume Retention" bigger than 28? Any higher
value is for nothing, because the tapes are already overwritten and you
cant get the data, so why storing volume information of data, which you
do not have anymore?
The same with "File -" and "Job Retention". Why store information about
jobs and files which cant be restored, 'cos the tapes have been already
overwritten?
Dont be to aggressive in this values - that you dont abolish datasets
which are still required.
If you adapt these values for your needs - look in the manual - you can
compact the database automatically.
One second reason is, that you maybe backup directories with a LOTS of
files - which obviously needs also a lots of entries in the database.
Check it, maybe you backup a news-dir (would be really crazy) or any
other dir, which contains a lots of files and does not have to be
backuped.
Another issue can be, that you use sqllite maybe - but I read that you
dump it with mysqldump. This idea is only for showing possible reasons
for it. ;-)) Dont beat me.
HTH,
Thomas
-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
