Amanda-Users

Re: holding disk via nfs?

2002-12-19 12:33:48
Subject: Re: holding disk via nfs?
From: Joshua Baker-LePain <jlb17 AT duke DOT edu>
To: Matthew Boeckman <matthewb AT saepio DOT com>
Date: Thu, 19 Dec 2002 11:48:15 -0500 (EST)
On Thu, 19 Dec 2002 at 10:10am, Matthew Boeckman wrote

> Running amanda 2.4.2, and have configured for tapeless operation to 
> spool backups to the holding disk. The holding disk is actually an NFS 
> mounted volume on a SNAP server. Problem I'm having is how to define 
> that 'partition'

Bleck.  But that's not important right now.

> the mount is 192.168.0.15:/amt_dump
> 
> Normally I would say something like:
> holdingdisk  amt_dump {
>       comment "blah"
>       directory "/bkup/amanda"
>       use 25 Gb
>       chunksize 15Gb
> 
> Which seems fine to me. What's happening is that lvl 0's on my biggest 
> partition are failing:
>    amt        /dev/sda9 lev 0 FAILED [dump larger than tape, but cannot 
> incremental dump new disk]
> 
> Here's the disk usage for that partition:
> /dev/sda9             122G  4.7G  111G   5%
> 
> There are 188G free on that holding disk! Why is amanda whining about this?

Doing this with 2.4.2, you still have to define a length to the tapes 
you're (not) using.  So, in your tapetype, "length" must be less than 
4.7G, which is why you're getting that error.  You could just crank it up 
to the size of your holding disk, or...

> Will this magically go away if I upgrade to 2.4.3 and use the file: driver?
> 
Well, not magically.  You'll need to set things up correctly.  But I'd 
highly recommend this, as 2.4.3 has this feature designed in.

-- 
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University


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