[OpenAFS] Strange Volume / Version Compatibility

Carsten Tolkmit ctolkmit@tng.de
Tue, 05 Feb 2002 12:25:06 +0100


Hi,

I am running SuSE 7.3 with a self-compiled kernel version 2.4.17 on the 
server, unchanged SuSE 7.3 on the client. I was just dumping a volume 
(by use of 'vos dump') and while it was dumped, the original volume on 
the server afs1 was corrupted.

I restored the volume from an old backup to server afs2 (SuSE7.3/2.4.17) 
as the salvage command on the volume told me sth. about 22500 orphaned 
files (arg!). I made a new mount point to the new volume. I have it 
accessible now, but:

when I do a 'vos listvol afs1' I get:

[...dadada...]
**** Could not attach volume 536870944 ****

Which is my old (corrupted) volume, of which I am pretty sure. I tried a 
'vos remove 536870944' which told me the volume has to be salvaged, so I 
did a 'bos salvage afs1 a 536870944 -showlog' which told me now:

 >>>
Starting salvage.
bos: salvage completed
SalvageLog:
@(#) OpenAFS 1.2.2 built  2002-01-14
02/05/2002 12:22:43 STARTING AFS SALVAGER 2.4 (/usr/afs/bin/salvager 
/vicepa 536870944)
02/05/2002 12:22:43 No applicable vice inodes on vicepa; not salvaged
Temporary file /vicepa/salvage.inodes.vicepa.432 is missing...
<<<

This happens every time I try to salvage it. If I give a different 
tmpdir using the -tmpdir switch, the same happens.

I even can't 'vos zap' the volume because it needs to be salvaged.

Any ideas why I keep losing my data?

Any idea how I can get rid of volume #536870944?

And last not least: Does the new OpenAFS 1.2.3 interoperate smoothly 
with OpenAFS1.2.2a servers and clients?

Thanks and regards,
Carsten