Jon LaBadie writes:
>On Mon, Apr 10, 2017 at 09:05:32AM -0400, [email protected] wrote:
>> Greetings.
>> 
>> I seem to be getting an awful lot of full (level-0) backups:
>> 
>Not unusual.  It a product of having a couple of large DLEs and several,
>relatively small DLEs.  Amanda tries to schedule so that each runs
>backup is about the same size.  The large ones skew the average needed
>each run and the only way to get nearer the average is to promote the
>small DLEs to level 0.

Makes sense, and yes, that pretty-much matches my setup.

I'm trying to figure out a way to split my two largest DLEs (one is
100GB and the other is about 60GB) into several smaller ones.

The problem is, as was pointed out to me earlier, I can't (or at least,
I don't know how to) get a good reference for "Based on your configuration,
this is what will be backed up tonight."  The only thing I can do is
run it, and when it's finished, check out what was backed up.

The other concern is that I'm using a hard disk as virtual tapes.
If I had 200 real tapes, it wouldn't matter how much space was used,
because they would get tossed in the vault until next time.  With a
hard disk, I can "overbook" my backups (I can define my changer as
"200 x 100GB disks" without actually needing 20TB, because most of my
backups aren't 100GB), but every GB I use on an unnecessary level-0
backup is a GB I don't have available when I need a "real" level-0 backup.

>Maybe a feature request could be an alternate
>scheduling scheme or a DLE setting to minimize promotions.

That would be nice.  In the meantime, thanks for the advice.

--hymie!     http://lactose.homelinux.net/~hymie    [email protected]

Reply via email to