Coda File System

Re: Doiung the repair limbo

From: Jan Harkes <>
Date: Mon, 9 Dec 2002 12:40:46 -0500
On Mon, Dec 09, 2002 at 06:34:45PM +0100, Steffen Neumann wrote:
> > You are right that this is a very unusual state, in fact we shouldn't be
> > able to enter Connected state while there are pending CML entries, and
> > you shouldn't have conflicts when everything is reintegrated. Can you do
> > 'cfs ck' to snapshot a checkpoint file. Maybe it will give an error
> Gaspra(sneumann):etc>cfs ck
> VIOC_CHECKPOINTML: Invalid argument

Ok, that should occur when there is no known CML for the volume. So in
that respect the listvol info is correct and there are no pending CML

> > (i.e. there is no CML). Or it will actually manage to write the .cml and
>   175 Dec  2 09:49
>  2048 Dec  2 09:49
> Nothing special in there
> > .tar files, interesting to know where that ows.c file really is located.
> I did a find /coda and locate and found 
> Gaspra:/coda # find /coda -noleaf -name *ows.c
> /coda/vol/kernel/src/linux-2.4.20/arch/sparc/kernel/windows.c
> /coda/vol/kernel/src/linux-2.4.20/arch/cris/kernel/shadows.c

Unlikely candidates, their location is so different, they are probably
even in a different volume.

>   Connection State is WriteDisconnected
> 	Gaspra:/coda/vol/ai/share # touch bla
> 	succeeded
> different machine(!):
> 	aipc10(sneumann):share>ls -l 
> 	-rw-r--r--   1 codadmin nogroup         0 Dec  9 18:31 bla
> again gaspra:
> 	Gaspra:/coda/vol/ai/share # cfs reconnect
> 	-rw-r--r--   1 codadmin nogroup         0 Dec  9 18:31 bla
> So the file appeared on both machines immediately.
> State is still strange...

Yeah, something is strange, when the client is write-disconnected,
updates shouldn't just show up everywhere at the same time. Reinit is
probably your best bet right now, I don't think we can figure out
exactly how we got into this state, but it definitely has something to
do with the repair session.

Received on 2002-12-09 12:41:55