oplock_break problem (PR#20285)

[Kirk Reiser]

  Jeremy Allison <jallison at cthulhu.engr.sgi.com> writes:

> The problem is that no one has ever come up with a *reproducible*
> test case - the clients just seem to "randomly" fail.

Would it be of any use to developers to have accounts on the
server/clients which generate the problems.

I suspect that it is probably possible to reproduce the problem if
more folks understood just how it all worked.  Of course, then we
could probably fix our own problems. 'grin'

It is possible though to provide an account or two on networks which
are incorrectly operating for an experienced person to poke around
with to find the problem.  It must be a fairly common misconception if
not a bug.  I suspect it has to do with my other problem which is I
don't seem to be able to move/copy more than one file across the
network.

  Kirk