On Thu, May 25, 2023 at 10:37:23AM +0200, Dumitru Ceara wrote: > Whenever an OpenFlow error is returned by OvS, trigger a reconnect of > the OpenFlow (rconn) connection. This will clear any installed OpenFlow > rules/groups. To ensure consistency, trigger a full I-P recompute too. > > An example of scenario that can result in an OpenFlow error returned by > OvS follows (describing two main loop iterations in ovn-controller): > - Iteration I: > a. get updates from SB > b. process these updates and generate "desired" openflows (lets assume > this generates quite a lot of desired openflow modifications) > c.1. add bundle-open msg to rconn > c.2. add openflow mod msgs to rconn (only some of these make it through, > the rest gets queued, the rconn is backlogged at this point). > c.3. add bundle-commit msg to rconn (this gets queued) > > - Iteration II: > a. get updates from SB (rconn is still backlogged) > b. process the updates and generate "desired" openflows (lets assume > this takes 10+ seconds for the specific SB updates) > > At some point, while step II.b was being executed OvS declared the bundle > operation (started at I.c.1) timeout. We now act on this error by > reconnecting which in turn triggers a flush of the rconn backlog and > gives more chance to the next full recompute to succeed in installing > all flows. > > Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2134880 > Reported-by: François Rigault <fr...@amadeus.com> > CC: Ilya Maximets <i.maxim...@ovn.org> > Signed-off-by: Dumitru Ceara <dce...@redhat.com>
Reviewed-by: Simon Horman <simon.hor...@corigine.com> _______________________________________________ dev mailing list d...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-dev
