Amanda-Users

Re: amstatus: no estimate and disk was stranded on waitq

2007-07-24 10:21:54
Subject: Re: amstatus: no estimate and disk was stranded on waitq
From: Jon LaBadie <jon AT jgcomp DOT com>
To: amanda-users AT amanda DOT org
Date: Tue, 24 Jul 2007 10:17:04 -0400
On Tue, Jul 24, 2007 at 08:46:56AM +0200, Marc Muehlfeld wrote:
> Hello,
> 
> fedora schrieb:
> >Questions:
> >1) "No writable valid tape found". What does it mean?
> 
> There was no tape in your drive/changer or it was not writeable (write 
> protected, broken tape, no tape with valid label,...)
> 

Just an addition:  Your tapecycle IIRC was 14.  Another meaning of "no valid"
tape is none that have NOT been used in the last 14 "successful tapings"
(a tapecycle) of amdump.  So suppose your tapecycle is 14 and you really
have exactly 14 tapes in rotation.  Then one of them goes bad for any
reason whatsoever.  Then you can not tape anything since all remaining
13 tapes were used in the last 14 tapings.


> 
> 
> >3) "can't switch to incremental dump". What does it mean? (causing failed
> >for the dump summary "domin11 -/lib/mysql 0 FAILED......." )
> 
> Amanda can't do a full dump on your tape, because of to less space. And 
> without having a full dump, Amanda can't do an incremental dump. This could 
> happen if you only use one tape and the estimated size is compressed (if 
> you use compression) bigger than your tape. Or you use more tapes and you 
> don't use splitting, ...
> 

Without being able to tape, the entire full dump of that DLE must go to the
holding disk.  But if there is insufficient space on the holding disk, or
if the reserve parameter is set high to ensure incrementals get saved on
the holding disk, then there is no place to put a level 0 of that DLE,
so it fails.  It can't switch to incremental because there is no other
level 0 on which to base the incremental.


> 
> >4) "cannot overwrite active tape DailySet1*".  What does it mean?
> 
> Seams you have too less tapes configured for one dumpcycle. Dumpcycle is 
> the number of of days in a backup cycle (e.g. one week). Amanda tries to do 
> a full backup at least this often (e.g. once per week). When you have e.g. 
> runspercycle=7 (how often you let amanda do amdump in dumpcycle) and a 
> tapecycle of 6, then amanda had to overwrite the first tape again to do the 
> backup on the last day. And this is what the message means. You can't 
> overwrite the tape, because you could never do a restore then, because of 
> the overwritten first tape. Amanda prevent you doing that.
> 
> # man amanda.conf
> - > tapecycle
> This is calculated by multiplying the number of amdump runs per dump cycle 
> (runspercycle parameter) times the number of tapes used per run (runtapes 
> parameter). Typically two to four times this calculated number of tapes are 
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> in rotation.

Your parameters have just exactly one times this calculated number.



-- 
Jon H. LaBadie                  jon AT jgcomp DOT com
 JG Computing
 4455 Province Line Road        (609) 252-0159
 Princeton, NJ  08540-4322      (609) 683-7220 (fax)