On Wed, Apr 6, 2016 at 3:10 AM, Han Zhou <zhou...@gmail.com> wrote:
>
> On Mon, Apr 4, 2016 at 5:58 AM, Russell Bryant <russ...@ovn.org> wrote:
> > - Each <code>localnet</code> logical port is implemented as a
> pair of
> > - patch ports, one in the integration bridge, one in a different
> > - bridge, with the same
> <code>external-ids:ovn-localnet-port</code>
> > - value.
> > + Each <code>localnet</code> and <code>gateway</code> logical
> port
> > + is implemented as a pair of patch ports, one in the
> integration
> > + bridge, one in a different bridge, with the same
> > + <code>external-ids:ovn-localnet-port</code> value.
>
> Is this ovn-localnet-port a typo?
>
No. It's still using the same external-id value. I did that because
"gateway" ports share almost all of the code for localnet ports, so it was
just convenient.
I thought about changing the name to avoid confusion, but hadn't decided on
exactly what to change yet.
> > diff --git a/ovn/controller/ovn-controller.c
> b/ovn/controller/ovn-controller.c
> > index 6027011..9bcda0d 100644
> > --- a/ovn/controller/ovn-controller.c
> > +++ b/ovn/controller/ovn-controller.c
> > @@ -326,7 +326,10 @@ main(int argc, char *argv[])
> > }
> >
> > if (br_int) {
> > - patch_run(&ctx, br_int, &local_datapaths,
> &patched_datapaths);
> > + if (chassis_id) {
>
> Shall we print an error log if chassis_id is NULL?
>
That's a good idea. It's already a problem today if chassis_id is NULL,
but we aren't logging anything about it. I may add that as a separate
patch.
> Otherwise looks good to me. I haven't tested it yet, but it is pretty
> straightforward approach.
>
--
Russell Bryant
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
