Bacula-users

Re: [Bacula-users] Reserving devices for pools

2015-04-06 09:01:28
Subject: Re: [Bacula-users] Reserving devices for pools
From: Josh Fisher <jfisher AT pvct DOT com>
To: "Clark, Patricia A." <clarkpa AT ornl DOT gov>, bacula-users <bacula-users AT lists.sourceforge DOT net>
Date: Mon, 06 Apr 2015 08:56:25 -0400
On 4/2/2015 1:23 PM, Clark, Patricia A. wrote:
> I have a tape library with several LTO5 tape drives.  I have 3 tape pools and 
> a scratch pool.  Pool1 is quite large and has most of the jobs.  The jobs are 
> in a fairly standard, monthly FULL, weekly DIFFERENTIAL, and daily 
> INCREMENTAL schedule.  The FULL backups are fairly evenly divided across 
> 1st-4th Saturdays and are quite large (multiple TBs).  Pools 2 & 3 are 
> smaller as are the jobs.  The issue that I am trying to solve is to have at 
> least one tape drive available for Pools 2 & 3 that Pool1 doesn't use.  It 
> appears that the only way to do this is to use a different media type for 
> Pools 2 & 3 and for any of the tape drives that I want reserved.  Of course 
> this would mean that Pools 2 & 3 would not be able to share the Scratch pool? 
>  Is anyone using multiple scratch pools or is that possible?

Yes, it is. I use something like the below to implement a second pool 
specific to the "full" pool that contains all volumes used for full 
backups. A scratch pool is not dependent on Media Type. In fact, any 
pool can contain volumes with different Media Types. The Media Type for 
a volume is taken from the Storage resource. I believe that it is taken 
from the Storage resource used to initially label the volume and that 
Media Type is completely independent of Pool.

I think separate scratch pools should allow what you are after, although 
you may also want a separate scratch pool for Differentials as well as 
Fulls. I ran into this issue when I noticed that with a single Scratch 
pool, if the storage where incrementals were being written runs out of 
space, then it will very quickly move all volumes from the Scratch pool 
into the Incremental pool and mark them in error due to the out of space 
condition. This effectively "uses" up all of the volumes, preventing a 
subsequent full job from running even if full volumes are being written 
to a different storage device that has plenty of space.

# A scratch pool specific to the "full" pool
Pool {
   Name = fullScratch
   Pool Type = Backup
   Recycle = yes
   AutoPrune = yes
   ActionOnPurge = Truncate
   RecyclePool = fullScratch
}

# The full pool that uses the "fullScratch" pool as its scratch pool
pool {
   Name = "full"
   Pool Type = Backup
   Recycle = yes
   RecyclePool = fullScratch
   ScratchPool = fullScratch    # the default is "Scratch"
   AutoPrune = yes
   ActionOnPurge = Truncate
   ...
}

> Patti Clark
> Linux System Administrator
> R&D Systems Support Oak Ridge National Laboratory
>

------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users

<Prev in Thread] Current Thread [Next in Thread>