On 8/20/2010 2:50 AM, Alex2398 wrote:
> Hi everyone!
> I have get the following issue while trying to make a weekly backup of each 
> client (Windows Client) and using one pool and one volume for each other. The 
> idea is having just one weekly file per backup and per client, trying to save 
> disk space.

If you are doing a full backup every week, you'll probably want to keep 
the old backups for 8 days.  That way, if the new backup fails, you 
still have the old backup for 1 day... just in case.

Personally, I'd be keeping backups longer than that. I think that 
relying upon just one full backup is not the best practice.

> The trouble occurs when some client is offline at the backup time. Then the 
> jobs prunes the volume and recycles it, waiting for the new data to come, but 
> as it's offline, after 2 minutes, the job drops an error (obviously). As 
> result, the volume gets empty, deleting the previous data, so, you lose the 
> backup until a new one is done when the client is online.

I think you'll find that pruning is done at the end of the job, not the 
start.  If this is not the case, please post a job email showing this.

> Is there a way to check connection before launching the job? I guess 
> recycling and purging volume should be done AFTER the job is complete, not 
> before.

Yes.  Include a Run Before script in your job. In that script, verify 
the connection.  Then terminate with the right error code to allow / 
deny the reset of the job.

How to verify depends on your OS etc.  I'd say a simple ping test would 
be enough.

-- 
Dan Langille - http://langille.org/

------------------------------------------------------------------------------
This SF.net email is sponsored by 

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users