"dev" <dev-boun...@openvswitch.org> wrote on 05/23/2016 03:48:39 AM:
> From: "Liran Schour" <lir...@il.ibm.com> > To: "Ben Pfaff" <b...@ovn.org> > Cc: ovs-dev <dev@openvswitch.org> > Date: 05/23/2016 03:49 AM > Subject: Re: [ovs-dev] [PATCH monitor_cond V6 11/11] RFC OVN: Quick > implementation of conditional monitoring > Sent by: "dev" <dev-boun...@openvswitch.org> > > Liran Schour/Haifa/IBM@IBMIL wrote on 17/05/2016 05:22:33 PM: > > > Conditional monitor of: Port_Binding, Logical_Flow, Multicast_Group > > MAC_Binding tables. As a result ovn-controller will be notified only > about > > records belongs to a datapath that is being served by this hypervisor. > > > > Hack: Ideally, logical datapath ID should be retrieved from Port_Binding > table > > and by that conditions should be composed only from logical datapath > IDs. > > In the meantime we first add the logical port to the conditions and > > on retrieval > > of port binding record, we add the logical datapath to the conditions. > > > > ovn-bridge-mappings unit test fails due to the fact that patch ports > will not > > be created on local bridge if there is no port binded to the logical > > datapath on > > the local host. > > > > Signed-off-by: Liran Schour <lir...@il.ibm.com> > > Pushed a tiny fix to this patch to pass ovn-bridge-mappings unit test. > Updated code is pushed to: https://github.com/liranschour/ovs.git branch: > monitor_cond_ovn. That patch *almost* gets us the rest of the way there - that test case switches between the current database and an empty database in lines 134-136 as a test of the database switching code - unfortunately, it looks like the patch ports aren't being recreated. That has the potential to cause headaches later on... Ryan _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
