On Monday 10 September 2007 13:50:16 Martin van Es wrote:
> On Monday 10 September 2007 13:33:58 you wrote:
> > Hi Martin,
> >
> > just a shot in the dark:
> >
> > do you repair as the same local uid who created the CMLs leading to
> > the conflict in the first hand?
> >
> > I think repair can become confused otherwise.
>
> Yes, as far as I can see I do.
>
> The expanded 'cached'-dir however is owned by root/nogroup on beginrepair,
> on my other client I can check that all files belong to user martin/nogroup
> (uid 1000 as the first user on both systems). The coda uid is 1000 as well.
>
> On the other hand I tried to repair as root before I wrote to the list
> because I suspected rights problems in the first place, but that didn't
> help anything.

I tried cfs flushcache, flushobject and others, but that didn't help. I 
(probably) solved the inconsistancy by (re)starting venus with -init.

The volume was in a very strange state btw: there were no pending changes but 
the inconsistent directory had the _localhost and [server] directory as 
normal readable entries without beginrepair and the volume was not writeable, 
but showed as read/write in cfs lv. The good client never had these entries.

So, all's well that ends well but I still have never been able to do a proper 
repair. Is repair working or a tool so much in development that it isn't 
deemed to do anything at all at the moment?

Best regards,
Martin