[OpenAFS] Duplicate VLDB entries

[Jeff Quinn]

jax > vos listvol afsfs10 vicepa
Total number of volumes on server afsfs10 partition /vicepa: 0 

Total volumes onLine 0 ; Total volumes offLine 0 ; Total busy 0

jax > vos syncvldb -server afsfs10 -part vicepa
VLDB synchronized with state of server afsfs10 partition /vicepa

jax > vos syncserv -server afsfs10 -part vicepa
Server afsfs10 partition /vicepa synchronized with VLDB

jax > vos listvldb -server afsfs10 -partition vicepa
...
Total entries: 1988

Seems like those commands aren't doing much to help. Any other ideas?

Thanks again,
-Jeff

-----Original Message-----
From: Stephen Joyce [mailto:stephen@physics.unc.edu] 
Sent: Wednesday, April 09, 2008 3:11 PM
To: Jeff Quinn
Subject: Re: [OpenAFS] Duplicate VLDB entries

Try

vos syncvldb -server <server> -part <part>
vos syncserv -server <server> -part <part>

If any of the volumes are replicated, you'd want to sync all the 
servers/parts which hold replicas too. Be sure to perform all the syncvldb 
commands before the syncserv commands.

In the steady- (non-problem) state, these commands should be 
non-destructive and do nothing, otherwise, they will change info/volumes as 
problems are corrected.

Cheers, Stephen
--
Stephen Joyce
Systems Administrator                                            P A N I C
Physics & Astronomy Department                         Physics & Astronomy
University of North Carolina at Chapel Hill         Network Infrastructure
voice: (919) 962-7214                                        and Computing
fax: (919) 962-0480                               http://www.panic.unc.edu

   "Lazy Programmers know that if a thing is worth doing, it's worth
   doing well -- unless doing it well takes so long that isn't worth
   doing any more. Then you just do it 'good enough'"
                  --- Programming Perl, p 282.

On Wed, 9 Apr 2008, Jeff Quinn wrote:

> We recently lost a partition on one of our servers and restored from tape.
> The volumes were all recovered, but when we did a vos exam on any of the
> restored volumes, it would have 2 identical entries in the vldb:
>
>
>
>    number of sites -> 2
>
>       server afsfs10.cl.msu.edu partition /vicepa RW Site
>
>       server afsfs10.cl.msu.edu partition /vicepa RW Site
>
>
>
> After moving the volume:
>
>
>
>    number of sites -> 2
>
>       server afsfs5.cl.msu.edu partition /vicepa RW Site
>
>       server afsfs10.cl.msu.edu partition /vicepa RW Site
>
>
>
> We moved all of the volumes off of that partition, so now they all look
like
> above.   A vos listvol of afsfs10 vicepa returns 0 volumes, but a vos
> listvldb returns all 1989 of them.    We then tried to do a vos syncvldb
> with a volume.  It reports that it has been synced, but there are still 2
> entries in the vldb.  We have also tried vos syncserv with the whole
> partition with the same result.
>
>
>
> We have also tried to manually remove an entry from the vldb using vos
> delentry, but you cannot specify a server/partition AND a volume.  A vos
> delentry on the volume removed both entries from the vldb.
>
>
>
> All volumes are accessible via their mount points.
>
>
>
> Any ideas on how to remedy the vldb discrepancy?
>
>
>
> Thanks,
>
> -Jeff Quinn
>
>