[OpenAFS] OpenAFS Backup Trouble (volume loss) on SuSE Linux 7.3 / Kernel 2.4.18

[Derrick J Brashear]

On Mon, 22 Apr 2002, Carsten Tolkmit wrote:

> > I assume this to be the infamous "CopyOnWrite" bug, but they'll have to
> > look in their FileLogs to confirm it.
> 
> 
> Ok - the following entries look interesting (to me, that is;-)):
> 
> Mon Apr 22 01:07:02 2002 CB: RCallBackConnectBack (host.c) failed for 
> host 213.178.69.42:7001
> Mon Apr 22 09:02:27 2002 CopyOnWrite failed: volume 536871180 in 
> partition /vicepa  (tried reading 8192, read 0, wrote 0, errno 2) volume 
> needs salvage
> 
> The first host mentioned (69.42) was most probably offline during that 
> time.

Which isn't interesting

> The second is the CopyOnWrite message.
> 
> Neither of the two occurs during backup though. Backup is running from 
> 22:55 to 23:55 (approx.)

The important thing is the second, which occurs after the backup. Based on
reports it only happens with the namei fileserver (the one you'll be using
under Linux) with either pthreads or lwp. It's being actively debugged,
but no answers yet. This is apparently a problem with both IBM AFS and
OpenAFS.