oplock tests - interesting!

[Jeremy Allison]

tridge at linuxcare.com wrote:
> 
> Jeremy,
> 
> Interesting news!
> 
> When a NT4 server fails to get a oplock break reply for a self break
> from smbtorture it waits 45 seconds then _continues_ as though the
> break had come in. It does not resend the break and does not drop the
> connection.

Yes, but this is actually a registry configuration
variable (whether to fail open or closed). Check out
the registry variable :

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\
LanmanServer\Parameters\EnableOplockForceClose

By default this is set to 1. This has the following
effect : "This value (1) will force closed the open 
instance of the client that has the oplock, risking
the loss of cached data".

Setting this to 0 will fail the second open.

Jeremy.

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