Veritas-bu

[Veritas-bu] error 41

2007-12-10 09:20:54
Subject: [Veritas-bu] error 41
From: "Fredrik" <fredrik.dahlberg AT rps.police DOT se>
To: <VERITAS-BU AT mailman.eng.auburn DOT edu>
Date: Mon, 10 Dec 2007 15:03:26 +0100

We have encountered error 41 when we changed a NW cluster. The new cluster is configured as the old but a error 41 is making my life a dark place.

We have a master and 3 media servers and backing it up to LTO3 tape drivers located in a SL8500.

Please advice.

 

Fredrik Dahlberg  - Backupadministrator
Tel dir: +46 (0)8 - 401 99 65
Tel mob: +46 (0)733-31 53 29
E-mail: fredrik.dahlberg AT rps.police DOT se


Från: veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] För jim fred
Skickat: den 10 december 2007 02:44
Till: VERITAS-BU AT mailman.eng.auburn DOT edu
Ämne: Re: [Veritas-bu] Reclaiming tapes to the scratch volume pool

 

Hi

You will also see this when you have decommissioned a media server.  Tapes belonging to the media server don't tend to be returned to the SCRATCH pool.   This when you get a  lot of tapes with DBBACKUP status when you run available_media.  The available_media script when it doesn't know what to do with a tape gives it a DBBACKUP status.  

 

This should fix that and other similar stuff.  
1. First generate a list of suspect tapes.  First column is the mediaIDs.
2. Run this script - it checks to see if their a images on each media and spews out a very simple output.  Patience is required it may take a will to run
------------------------------------------------------------------------
#!/bin/ksh
# Author Jim McDonald , Sydney Aust 
# usage : script name followed by a file that has one column with media IDs to check
# output goes to /tmp/med-in-images and genrates rows of
#     media_id  FRAG  client_name
#     media_id  IMAG  client_name
#     media_id  no entity was found
# sorted by media_id
# "no entity was found" = no images on that tape
TMP_FILE=/tmp/med-in-images
> $TMP_FILE
let CNT=0
for MED in $(cat $1 | awk '{ print $1 }'); do
 let CNT=CNT+1
 echo $CNT , $MED  # let you know its doing some work
  bpimmedia -mediaid $MED 2>&1 | nawk  -v media=$MED '
            $1 == "FRAG" { printf "%s  FRAG %s \n",media,$10}
            $1 == "IMAG" { printf "%s  IMAG %s \n",media,$2}
           /found/  { printf "%s  %s\n",media,$0}
          '  >> $TMP_FILE
done
------------------------------------------------------------------------------------------------------------
3. You now have a file that list what is on media, incuded those that have nothing.
4. Create a list of files that you want to fix,  Run
    grep "no entity was found" /tmp/med-in-images  > to_deassign
5. Run this script to
------------------------------------------------------------------------------------------------------------
#!/bin/ksh
# Author Jim McDonald , Sydney Aust 
# this generates lines like :
# vmquery -deassignbyid AZ0070 3
# input : a file with one column of media  IDs
METHOD="vmquery -deassignbyid"
for MEDIA in $(cat $1 | awk '{ print $1 }'); do
  vmquery -m $MEDIA 2>&1 | sed s/0x//g | tr -d ")" | tr -d "(" | \
    nawk -v how="$METHOD" ' /media ID:/   { printf"\n%s %s ",how,$3}
                                        /volume pool:/   { printf"%s ",$4}
                                        /status:/           { printf"%s",$2 } '
done
--------------------------------------------------------------------------------------------------------------
METHOD can be edited to include sudo if required


This generates outout for each media.....eg
vmquery -deassignbyid Z00471 38 0
vmquery -deassignbyid CV1700 1 1
vmquery -deassignbyid CV1905 1 1
vmquery -deassignbyid CV0901 25 0

from their it simple to deassign all the media i.e your SCRATCH pool should automagically grow

Jim

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu