On Fri, 23 Feb 2001, Jan Harkes wrote:

> 
> The X CML entries remaining indicates that the client was unable to
> sends all updates back to the server. There is possibly an associated
> error message in /usr/coda/etc/console. The object/operation that failed
> reintegration is the first one in /vice/spool/<uid>/<volumeY>.cml.
> 
> Sometimes the kernel is quite persistent in keeping stale directory
> contents around, especially if an application is still holding a
> reference to a dentry. Try to stop all programs/shells that might be
> running with their current working directory somewhere in /coda.
> 
> Then walk back up the tree to where you were working, one of the parent
> directories might be marked in-conflict, but wasn't expanded because of
> those refcounts.
> 
> Jan
> 

Well, O.K., those are good clues. Unfortunately, the directory
/vice/spool on the server is empty. I tried killing all processes that
might be attached to coda, and when I logged in again, I found that
the number of CML entries was down from 16 to 12, so it seems like
something happened, but I am still stuck. I also tried looking for conflicts
in some of the leaf level directories I was working in, but didn't find
any. However, there was a long series of messages in /usr/coda/etc/console
saying that my volume has unrepaired local subtrees. Is there any systematic
way I can use to find out what needs to be repaired?

	Ken