I have the same question in 3 different formats. 

1. HELP!!


2. I'm trying to setup a system where I start with one client and add
more as needed and where I start with one disk as storage and add more
as needed. Now the actual storage disks will be just JBOD but they will
be always mounted into the filesystem of the storage server. I do not
want to have to create any raid system for the storage. If a storage
disk is going bad I will simply replace it and delete all the
jobs/volumes that was stored on that disk. I'm counting on bacula to
realize what clients now needs to do a full/differential backup to again
have a valid backup. 

So what would be a good configuration for this setup??


3. I have been testing bacula for a few days now and while basic backup
of one client works fine in the different setup I tried it's hard to
figure out a way that is actually extensible when trying to store to
disks. 

Clearly mapping one disk to one volume is not going to work very well
due to retention time and reuse.

creating a single disk storage and then having the clients do a  singe
job to single volume worked nicely and is using the storage efficiently
when using one disk but I never figured out how to extend this system to
two disks. there is a one to one mapping from job to storage and I need
more than one storage device when adding the second disk. adding the
storage was just duplication the first one but how do I then use it ???
I could assign a new client to the disk but this would really badly hurt
the utilization of the disk. very inefficient. 

I tried to experiment with the autoloader feature but then the meaning
of "Archive Device" changed meaning from the directory where volumes is
created in into the actual volume totally breaking the functionality.

So now I'm thinking about some system where I have one storage device
per disk. But instead of allowing bacula to create the volumes at
demand(cant make it work on multiple storage devices) doing it manually
and putting the volumes in one common pool. by creating relatively small
volumes I can increase the probability that the data on every volume
will be from the same job and will have the same retention period so the
volume can efficiently be reused. Would this work ?? 


------------------------------------------------------------------------------
Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
Finally, a world-class log management solution at an even better price-free!
Download using promo code Free_Logger_4_Dev2Dev. Offer expires 
February 28th, so secure your free ArcSight Logger TODAY! 
http://p.sf.net/sfu/arcsight-sfd2d
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users