[PATCH] Fix bug #13121 - Non-smbd processes using kernel oplocks can hang smbd
Jeremy Allison
jra at samba.org
Wed Nov 29 03:21:51 UTC 2017
On Wed, Nov 29, 2017 at 04:05:43PM +1300, Andrew Bartlett wrote:
> On Tue, 2017-11-28 at 17:27 -0800, Jeremy Allison via samba-technical
> wrote:
> > On Wed, Nov 29, 2017 at 02:24:36PM +1300, Gary Lockyer wrote:
> > > Jeremy,
> > > Can we put it into flapping until then, as it's breaking at least
> > > 50% of our cloud builds
> >
> > Can this be a private patch ?
>
> Sadly no. A local patch carried across all our team members will end
> up in master sooner rather than later because it will have to be in all
> the development trees, all the time. As you know, running autobuild in
> Samba 14.04 VMs is central to our workflow and is central to ensuring
> we don't push our own flapping tests to master (we run them a lot!).
>
> I would rather this was officially acknowledged until it is fixed.
>
> Testing beyond the specific sn-devel machine is also key to my efforts
> to make pre-push Samba testing more accessible.
>
> > It isn't breaking anywhere else
> > than your cloud builds and I'd like to know why that is rather
> > than ignoring it. What's different about your cloud kernel ?
>
> It is the official kernel from Ubuntu, it isn't Catalyst specific.
>
> > That's the only thing I can think of that would cause the
> > process to not be sent a signal.
>
> I can get you a login on such a VM if you like.
I can take a look, but right now everything is pointing to
a problem in your cloud environment. Gary, you said you'd
been able to reproduce it - was that on a local build ?
What kernel was it ?
More information about the samba-technical
mailing list