Thanks for the info. In my case I’ll be backing up to disk. Offsite storage will be done using a combination of things. Not sure how it all works yet but here’s what was suggested so far.
1.
Use version 7
2.
Run a base job and copy that to the remote server
3.
Use a virtual job to convert incremental backups into a full backup
4.
Remote servers using copy jobs to transfer data offsite – still looking into this one.
From: Josh Fisher [mailto:jfisher AT pvct DOT com]
Sent: Tuesday, January 27, 2015 4:28 AM
To: bacula-users AT lists.sourceforge DOT net
Subject: Re: [Bacula-users] Should I have a server in every location?
On 1/26/2015 10:00 PM, Damien Hull wrote:
I’m brand new to Bacula. Getting some good information from the list. As I’ve mentioned before, I’ve got three offices in different locations. I’m needing offsite storage of data as part of a disaster recovery plan. I’ve got a couple of
questions
·
Should I have a server in every location?
·
Can Bacula servers send data between each other?
Here’s my thinking
1.
Local backup in case someone deletes a file
For just restoring a deleted file on occasion, local backup is not needed. Only the restored file(s) will be transmitted. For restoring entire machines, a local backup will be much faster.
Whether or not to have local servers depends on several things. In addition to the expensive of separate tape hardware, tapes, etc. at every location, it is an administrative headache. Tapes must be handled and stored at each location. An operator is needed
at each location to load and change tapes and to manage sending tapes to the offsite location, etc. It may mean more people have to be trusted with sensitive data. Branch to main office network bandwidth must be considered. Local servers may be the only choice
if bandwidth is limited. On the other hand, centralized backup may be the only choice if the remote sites do not have the personnel. So it really depends on which scenario will be the best fit for your organization.
2.
Use another remote server for offsite storage
It is possible. Backup jobs on local volumes can be migrated to volumes at the main site over the wire. It of course requires the same bandwidth as if the job were written to a main site volume directly, but the migration can be scheduled for off hours and
writing to local volumes allows client jobs to finish much faster.
Thanks!
