[OpenAFS] Re: Volser: GetVolInfo: Could not attach volume (/vicepa:) error=101

Andrew Deason adeason@sinenomine.net
Tue, 31 Jan 2012 11:30:32 -0600


On Mon, 30 Jan 2012 15:22:34 -0500
Dan Scott <danieljamesscott@gmail.com> wrote:

> > Do FileLog or SalvageLog mention that volume? (either by id number
> > or by name)
> 
> FileLog doesn't. The last entry just shows:
> 
> Mon Jan 30 15:06:08 2012 File Server started Mon Jan 30 15:06:08 2012
> 
> The SalvageLog shows salvage entries for all volumes on the partition
> (including the one which gave an error):

Okay, so what about 'vos examine' ? Or does that still just give an
error that it can't attach the volume?

> > Also try running 'fs checks', 'fs checkv', and 'fs flushm
> > /afs/example.com/volume' on the client, and see if that makes any
> > difference.
> 
> djscott@pc35:~$ fs checks
> All servers are running.
> djscott@pc35:~$ fs checkv
> All volumeID/name mappings checked.
> djscott@pc35:~$
> 
> The flushm didn't help. Still saying 'No such device'.

How busy are the client and fileserver here? Can you get debug
information from them? There's a few things that could be wrong, so it
can be easier to just look more closely at the error than just checking
random stuff. If the client is not busy doing other things in AFS, run
this as root:

fstrace clear cm
fstrace setlog cmfx -buffers 200
fstrace sets cm -active
ls /afs/example.com/volume # should give 'No such device'
fstrace dump cm > /tmp/fstrace.log
fstrace sets cm -inactive

And put the fstrace.log file somewhere where I can look at it. That will
contain some information about your cell, like volume names, and
possibly filenames, IP addresses, etc. So if you're not comfortable with
that, then we can try other stuff.

-- 
Andrew Deason
adeason@sinenomine.net