Bacula-users

Re: [Bacula-users] backup to multiple disks

2013-05-03 11:07:50
Subject: Re: [Bacula-users] backup to multiple disks
From: Radosław Korzeniewski <radoslaw AT korzeniewski DOT net>
To: Dimitri Maziuk <dmaziuk AT bmrb.wisc DOT edu>
Date: Fri, 3 May 2013 17:04:17 +0200
Hello,

2013/5/2 Dimitri Maziuk <dmaziuk AT bmrb.wisc DOT edu>
Hi everyone,

this must be a FAQ but neither TFM nor google are being very helpful...

I have a storage server with 12 SATA drives on an HBA. I want to backup
to those drives, using file volumes of 50GB or less. The idea is to
manually swap the drives as they get full.

Drop this idea. If you need to move your backups out of datacenter then buy a tape library.

If this is not a case then why to do something manually?

Why not to use LVM or RAID to use all available hard disk space (12 x sata) for backups? Believe me, it is simpler to create a single filesystem for backups then use vchanger script (it is overcomplicated and IMVHO useless). When you have a single filesystem for backups then you can configure SD something like:

Autochanger {
  Name = disk-storage
  Device = File-Storage-xx
  (...)
  Changer Command = ""
  Changer Device = "/dev/null"
}
 
And a number of Device resources which define a required level of concurrency and points to the same Archive directory.


I'm looking for the implementation with least moving parts. From what I
see out there my choices are

1. Autochanger.

Can I just omit the "Changer Command =" (or set it to /bin/true) and
it'll magically work?

More or less, Yes... :) above.
 

Or do I have to have a "fake autochanger" script -- and if I do, how is
it supposed to work with 12 drives all mounted at once? E.g. TFM has an
example of 2-drive configuration, but it doesn't say what "mtx-chager
loaded" is supposed to return in that case.


You can use vchanger, disk-changer or your custom script for that purpose. My best advice is to follow a KISS way above.
 
2. Give each client its own disk. I'll consider it if all else fails,
but obviously this is far from optmal.

Yes, it is far from optimal.

best regards
--
Radosław Korzeniewski
radoslaw AT korzeniewski DOT net
------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite
It's a free troubleshooting tool designed for production
Get down to code-level detail for bottlenecks, with <2% overhead.
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap2
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users