On 9/19/2014 12:35 PM, Dimitri Maziuk wrote:
> On 09/19/2014 11:10 AM, Josh Fisher wrote:
>> On 9/19/2014 11:35 AM, Hendrik Visage wrote:
>>> The servers are big and oldish and they have enough disk space for my
>>> current needs and thus not sensible to buy extra hardware yet, but to
>>> have them all run 7x24x366 is a waste of electricity (and heat and
>>> and), so the idea is to turn on the one for that day's backups, and
>>> let it power off via cron after backups finished. If a recovery is
>>> needed, I can turn them all on for that recovery session.
>> If they both have sufficient disk space, then I suggest configuring the
>> two servers in a two-node Pacemaker cluster using DRBD shared storage.
> That would be OK if the goal was HA, but it's the opposite of saving
> electricity.
>
> If you have to turn them on manually, you might as well do one extra
> manual step and mount the filesystem. You could probably squeeze
> something out of round-robin dns if you really want to automate the
> mounting bit...

Except that you would have to turn them both on for a restore. The 
cluster (with only one node running at a time) prevents having to boot a 
server just to run a restore. If you don't do many restores, then it 
wouldn't be an issue. I have to restore files that users have "lost" all 
of the time, so for me losing the ability to do a restore from a remote 
location would be a deal breaker.


------------------------------------------------------------------------------
Slashdot TV.  Video for Nerds.  Stuff that Matters.
http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users