19:05:53 #startmeeting weekly ovsdb meeting 19:05:53 Meeting started Tue Jun 3 19:05:53 2014 UTC. The chair is Madhu. Information about MeetBot at http://ci.openstack.org/meetbot.html. 19:05:53 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:05:53 The meeting name has been set to 'weekly_ovsdb_meeting' 19:06:14 #topic Current status on OpenStack integration for Helium 19:07:50 #info biggest missing pieces in OVSDB for Hydrogen were OpenStack services 19:08:39 Will info’s show up from non-chaired folks? 19:09:13 tbachman: nope 19:09:28 but thanks for the sentiment 19:09:32 lol 19:09:45 * tbachman writes meetbot-ese into the ethersphere.... 19:10:59 tbachman: Ha! 19:15:05 hey guys, can't get on webex but I'm here 19:16:54 * mestery waves at Mierdin. 19:16:58 Mierdin: Hey there man! 19:20:32 Madhu: I think br-ex does end up getting configured for the OF10Provider, but not the OF13Provider. 19:21:13 tbachman: That would jive with what I have, as I think devstack defaults to OF13Provider. 19:21:42 can someone please take meeting notes guys ? :) 19:23:09 tbachman: Yes, #info shows up for non-chair folks. 19:23:22 dave_tucker: I thought they did? #links show up. 19:23:44 #chair dave_tucker mestery tbachman 19:23:44 Current chairs: Madhu dave_tucker mestery tbachman 19:23:44 mestery: hey! Sitting in NETCONF session at NANOG and also in OVSDB meeting in IRC. I think that's some kind of blasphemy. :) 19:23:54 there u go guys ;) 19:23:59 Madhu: thx! 19:23:59 :) 19:24:12 Mierdin: Course you kids and your fancy "Open Source." 19:24:17 Mierdin: Did you hear that at NANOG yet? :) 19:26:55 #info Floating IPs needs work, as br-ex isn’t getting configured (e.g. might need OF_NORMAL rule) when using OF13 plugin 19:28:07 #info LBaaS is being worked on by Chris Small (student at University of Washington) 19:28:40 mestery: qq on #link https://review.openstack.org/#/c/84223/ 19:29:39 #info Chris is working on a test version using Nicira extensions in OVS for LBaaS. 19:32:16 flaviof: Yes, please go ahead. 19:32:22 #info Madhu asks if OVS is doing load balancing for us, we need to have some sort of idea how nodes participate in a distributed way 19:33:16 mestery: thanks. looking @ that set of changes, is there ml2 re-factoring that would affect L3 fwd ? And if so, is this ice house/Junos timeframe? 19:33:33 mestery: wondering how it jives with Helium 19:33:51 #info chsmall says that essentially the idea would be that it would be done in distributed way by having the service calls go to a specific OVS/ODL configuration, which would be separated from what ML2 does, so that you can mix and match at the OpenStack layer 19:34:03 flaviof: That work only affects the L3 agent, not ML2 with the OVS agent, so no refactoring needed. 19:35:04 mestery: ic. Need to understand how l3 fwd will percolate into ODL (if not using ml2). Maybe we can talk offline when you get free cycles?!? 19:37:04 mestery: not really, though I suspect will get more of them when I do a lightning talk on ODL with Dave later today :) 19:37:12 #info TCP flags are in 19:37:40 #info we can now do flowmod tunnels 19:38:56 #info registries are a much more flexible pipeline — can do load, move, copy of any field. 19:39:33 #info when registries are in place, that will unblock ARP, and ARP response can move forward. 19:42:06 #info If you have any NXM extension needs for a feature, bring that up sooner rather than later, so that we can push for support in the OF plugin 19:43:58 #info shague__ is working on getting the existing HA proxy version of LBaaS working 19:49:22 #info http://pastebin.com/b59H0HU7 19:53:12 tbachman: oh, meaning it already works on 1.0 (heh, somehow i missed that) 19:53:25 yeah 19:53:36 cut-n-waste fails again 19:53:48 well — I should say that there’s code there that *should* make it work ;) 19:54:27 #info tbachman to verify that floating IPs don’t work for 1.3, will open a bug and fix for hydrogen-stable 19:56:09 #info Sudheendra Murthy from EBay is working with Ashwin for the schema changes 19:57:19 tbachman: don't thikn it's just one-liner 19:57:29 :( 19:57:36 maybe 2-liner ? ;) 19:58:26 1.3 initializeFlowRules(Node node, String bridgeName) is kinda busted 19:58:38 yeah — just looking into that now 19:59:10 It’s kind of special-cased :( 19:59:10 k 19:59:14 yup 19:59:15 wishful thinking 19:59:20 aka hardcoded 20:00:02 more than 2-liner 20:00:31 yup (2-liner was tongue-in-cheek) 20:00:41 :) 20:00:48 * tbachman is soooooo literal sometimes 20:00:49 lol 20:00:55 cdub: thx! 20:01:03 keepin' it real 20:01:07 lol 20:02:15 * cdub drops for next call 20:03:07 #info flaviof is looking at ML2 to understand how things are hooked up. 20:03:40 and finding bugs 20:03:43 Madhu: be sure to endmeeting :) 20:03:48 * cdub ^5s flaviof 20:03:59 #endmeeting