Jeremy Allison jallison at cthulhu.engr.sgi.com
Fri Sep 17 16:52:10 GMT 1999

tridge at linuxcare.com wrote:
> curiouser and curiouser!
> Jeremy, have a look at oplock_break6.dat. It shows that when NT4 gets
> a openX from a different client than the one that has a file oplocked
> it
> 1) sends a break request as expected
> 2) fails the open with ERRbadshare when the break times out
> 3) if a 2nd open comes immediately afterwards from the second client
>    then it succeeds!
> I can only attribute this to a NT4 oplock bug. Any other ideas on how
> to rationalise this?

It's the "fail closed" action. This is strange. One of
your earlier tests (the unlink one) implies that on break
timeout the oplock is dropped, but the file is left open.
This test implies that on break timeout the oplock is
dropped and the file is closed.

Any other explainations you can think of ?


Buying an operating system without source is like buying
a self-assembly Space Shuttle with no instructions.

More information about the samba-technical mailing list