High oplock break fail count

[Eric Sisler]

Martin Hoffesommer <mh at soft-enterprises.com> wrote:

>we've just recently switched to a new file server upgrading to samba 2.0.3.
>Unfortunately I've noticed that a couple of times a day a smbd process dies
>with 'oplock break failed' so I actually had to turn off oplocks thus
>degrading performance drastically. Does anybody have any idea why oplock
>breaks are failing so often now?

I'm in the same boat and unfortunately I don't have any answers for you.
We're running RedHat 5.2 and I've reverted to samba-1.9.18p10-52.4 because
of the oplock problem.

It's probably not very scientific, but a quick count of the number of times
"oplock" occured in the samba logfiles revealed the following:

samba 2.0.3, running for 1 week = 549 times
samba 1.9.18 running for 2 weeks = 37 times

Ouch!

Any ideas what's going on with oplocks?  I've even seen some oplock break
failures on read-only shares that have "fake oplocks = yes".

-Eric



Eric Sisler
Library Computer Technician
Westminster Public Library
Westminster, CO, USA
esisler at libris.ci.westminster.co.us

Linux - don't fear the Penguin.
Want to know what we use Linux for?
Visit http://gromit.ci.westminster.co.us/linux