On Fri, Mar 04, 2005 at 07:43:12PM -0800, Mitch (Bitblock) wrote:
> I verified the problem exists in the latest svn - can test again as needed -
> just let me know.
> If logs or anything would be helpful, can do that too - wasn't sure what to
> look - didn't want to send a bunch of useless data.

Is it perfectly reproducible ? If so, then 2 ethereal capture traces
of the problem occurring with oplocks turned on, and exactly the same
activity with oplocks turned off would help. Also if you can run it with
oplocks on to a W2K server and capture the trace that would be perfect.