I am seeing packet duplication using omping from a bare metal server to a VM running on a KVM host connected to the OVS. The amount of multicast packets that are getting duplicated is directly proportional to the number of VLAs on the trunk. If I were to prune the VLANs on the trunk the number of duplicates that are seen are reduced. The bare metal server is the one who is reporting duplicate packets. It looks like the OVS is not respecting VLAN tags for multicast packets and is just sending the traffic down all VLANs.
On Fri, Jun 24, 2016 at 12:47 PM Ben Pfaff <b...@ovn.org> wrote: > OK. > > What packets get duplicated, and in what circumstances? > > On Fri, Jun 24, 2016 at 12:55:53PM +0000, Dominick Krachtus (TT) wrote: > > I am seeing packet duplications on version 2.5.1. Version 2.5.1 dose > solve > > all of the other issues we were seeing. > > > > On Fri, Jun 24, 2016, 00:11 Ben Pfaff <b...@ovn.org> wrote: > > > > > You say that 2.5 fixed the problem, so in what version are you > observing > > > it? > > > > > > On Fri, Jun 24, 2016 at 12:39:02AM +0000, Dominick Krachtus (TT) wrote: > > > > It appears I have spoken too soon after more I am still seeing issues > > > with > > > > omping in some circumstances. When I use omping from a host somewhere > > > else > > > > on the same network to a VM on the KVM host I am seeing packets > getting > > > > duplicated. The duplicate packet issue was present in 2.4 and > earlier and > > > > was fixed in 2.5. > > > > > > > > On Wed, Jun 22, 2016, 15:05 Ben Pfaff <b...@ovn.org> wrote: > > > > > > > > > Hurray! Thanks for the update. > > > > > > > > > > On Wed, Jun 22, 2016 at 06:37:55PM +0000, Dominick Krachtus (TT) > wrote: > > > > > > Today I have built ovs from branch-2.5 which is version 2.5.1 > > > according > > > > > to > > > > > > ovs-vsctl -V. This version has solved all issues that we are > seeing > > > with > > > > > > multicast. There is no longer flooding into other VMs and the OVS > > > MCAST > > > > > > table is properly being updated when IGMP reports are sent out. > > > > > > > > > > > > ovs-vsctl (Open vSwitch) 2.5.1 > > > > > > > > > > > > Compiled Jun 22 2016 12:26:40 > > > > > > > > > > > > DB Schema 7.12.1 > > > > > > > > > > > > > > > > > > > > > > > > On Mon, May 9, 2016, 19:21 Ben Pfaff <b...@ovn.org> wrote: > > > > > > > > > > > > > On Thu, May 05, 2016 at 02:39:34PM +0000, Dominick Krachtus > (TT) > > > wrote: > > > > > > > > Below are the steps i have taken to setup our current testing > > > > > > > environment. > > > > > > > > As of now this environment is up and functioning with > multicast > > > > > working. > > > > > > > > There are however some oddities with multicast. Testing with > > > omping > > > > > > > > produces miked results while using our application works > without > > > > > issues. > > > > > > > So > > > > > > > > far I have tried four different ways of configuring the > > > environment > > > > > and > > > > > > > the > > > > > > > > current config has been working the best. > > > > > > > > > > > > > > > > I have upgraded the OVS to 2.5 and saw that it fixed a > majority > > > of > > > > > the > > > > > > > > ompoing issues but not the MCAST state table aging issue is > still > > > > > > > present. > > > > > > > > This appears to be some type of possible bug with the IGMP > > > snooping > > > > > > > > implementation within the OVS. > > > > > > > > > > > > > > I've recently posted a couple of patches for IGMP issues. > Please > > > try > > > > > > > them and see whether they fix the problems you are seeing. > > > > > > > > > > > > > > Patch 1: https://patchwork.ozlabs.org/patch/619681/ > > > > > > > Patch 2: > > > http://openvswitch.org/pipermail/discuss/2016-May/021234.html > > > > > > > > > > > > > > Thanks, > > > > > > > > > > > > > > Ben. > > > > > > > > > > > > > > > >
_______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss
