15:04:31 <yamahata> #startmeeting neutron_northbound 15:04:31 <odl_meetbot> Meeting started Wed Jul 27 15:04:31 2016 UTC. The chair is yamahata. Information about MeetBot at http://ci.openstack.org/meetbot.html. 15:04:31 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:04:31 <odl_meetbot> The meeting name has been set to 'neutron_northbound' 15:04:43 <yamahata> #chair vthapar mkolesni asomya 15:04:43 <odl_meetbot> Current chairs: asomya mkolesni vthapar yamahata 15:04:50 <yamahata> #topic agenda bashing and roll call 15:04:56 <yamahata> #info yamahata 15:05:02 <mkolesni> #info mkolesni 15:05:07 <asomya> #info asomya 15:05:13 <yamahata> #link https://wiki.opendaylight.org/view/NeutronNorthbound:Meetings meeting adgenda 15:05:19 <vthapar> #info vthapar 15:05:24 <yamahata> I updated the page. 15:05:57 <yamahata> In addition to usual topics, I newly added "migrating to new features" 15:06:22 <yamahata> I mean, what needs to be done to switch to new features like v2 driver for Newton cycle. 15:06:30 <yamahata> any other agenda items? 15:07:08 <mkolesni> what was decided about vlan aware vms? 15:07:17 <mkolesni> its pushed back to carbon? 15:07:28 <yamahata> it's for carbon. 15:07:31 <vthapar> yes 15:07:34 <mkolesni> ok 15:07:49 <yamahata> #topic Announcements 15:08:09 <yamahata> Boron M5 was past. Now it's code freeze. 15:08:24 <yamahata> #link https://wiki.opendaylight.org/view/NeutronNorthbound:Boron_Release_Plan 15:08:52 <yamahata> Soon the branch will be cut and master branch will be open for Carbon. 15:09:08 <yamahata> Personally I'll focus on networking-odl. 15:09:23 <yamahata> voting for openstack summit is open 15:09:32 <yamahata> #link https://www.openstack.org/summit/barcelona-2016/vote-for-speakers/ 15:09:48 <yamahata> So please vote for our and ODL proposals 15:09:57 <mkolesni> +1 15:09:59 <yamahata> any other announcement? 15:10:53 <yamahata> seems nothing, next. 15:10:55 <yamahata> #topic action items from last meeting 15:11:20 <yamahata> yamahata report M5 report, updated meeting page and decidec vlan-aware-vm plan. done 15:12:02 <yamahata> The next planned topic is "migrating to new features". before that I'd like to finish other items before. 15:12:13 <yamahata> #topic patches/bugs 15:12:31 <yamahata> I've uploaded a bunch of patches to enabled new features with devstack 15:12:53 <yamahata> My goal is to enable new features with openstack CI 15:13:31 <yamahata> Any patches/bugs which we need special attention for? 15:14:35 <yamahata> So far I hard from some complaining about logging of networking-odl is too verbose. 15:14:46 <yamahata> Anyway we will see by CI. 15:14:56 <mkolesni> there was some report by a guy from hpe 15:15:15 <mkolesni> but i think its low priority since he's talking about debug logs 15:15:51 <yamahata> #link https://bugs.launchpad.net/networking-odl/+bug/1605112 q-svc.log if filled with [networking_odl.journal.journal] printouts 15:16:27 <yamahata> #link https://bugs.launchpad.net/networking-odl/+bug/1604678 networking ODL-V1 VS. networking ODL-V2 - PortsAdminExtendedAttrsTestJSON test cases fail 15:16:43 <yamahata> I made 1605112 low priority 15:17:20 <yamahata> any volunteer for 1604678 15:17:38 <asomya> 1605112 might be invalid, i'll triage it 15:17:41 <yamahata> anyway after enabling CI, I expect more will pop up. 15:17:55 <yamahata> asomya: cool thanks. 15:18:49 <yamahata> any other bugs/patches? 15:19:38 <yamahata> okay we can move on to migration to new features. 15:19:42 <yamahata> #topic migrating to new features 15:20:30 <yamahata> Although there are many things to be added for new features, we should also think about how to switch to them. 15:20:39 <yamahata> I put some 15:20:46 <yamahata> #link https://wiki.opendaylight.org/view/NeutronNorthbound:Meetings 15:20:57 <yamahata> I'll create a wiki page for it 15:21:06 <yamahata> #action yamahata create a wiki page for feature migration 15:21:40 <yamahata> at least we should enable openstack CI and stabilize/harden them. 15:21:42 <mkolesni> we need to do it gradually 15:21:54 <mkolesni> i.e. first switch to v2 driver 15:22:07 <mkolesni> once its stable switch to vpnservice netvirt 15:22:20 <mkolesni> if we do all at once we wont know where the problems come from 15:22:53 <yamahata> mkolesni: fully agree. one thing at a time 15:23:24 <yamahata> vpnservice isn't feature parity for Boron. It will take for Carbon. So for newton cycle it will be experimental. 15:23:38 <yamahata> I suppose vpnservice switch will be the last. 15:23:51 <mkolesni> makes sense 15:24:38 <yamahata> For v2 driver, in addition to CI, what else should be done? 15:25:04 <mkolesni> once ci is stable we should make it default 15:25:21 <yamahata> mkolesni: +1 15:26:09 <yamahata> benchmark with many resources is necessary. e.g. 10k ports 15:26:30 <mkolesni> do we have h/w for that? 15:26:34 <yamahata> Probably Rally can be used. and we'd like to know latency increase and CPU usage. 15:26:59 <yamahata> we don't have any in public. 15:27:12 <mkolesni> i thought rally is for the api load i.e. running many requests 15:27:30 <mkolesni> but not necessarily with many resources 15:27:58 <yamahata> mkolesni: right. 15:28:18 <yamahata> We may have to create simple script for benchmark if rally isn't suitable. 15:28:24 <mkolesni> we need to see if its possible upstream 15:28:50 <yamahata> yeah. 15:28:50 <mkolesni> should be but well see.. 15:29:15 <mkolesni> probably we need tests for full sync 15:29:23 <mkolesni> and journal recovery 15:29:38 <mkolesni> and other maintenance features we have 15:29:56 <yamahata> new tests would be needed for those. 15:30:18 <mkolesni> i dont think its on tempest level 15:30:35 <mkolesni> probably we need some system test layer in the project 15:33:04 <yamahata> For pseudo agent port binding, it needs to be tested at first 15:33:04 <mkolesni> ill try to work on it but cant guarantee 15:33:12 <yamahata> mkolesni: cool. 15:33:33 <yamahata> I'm also trying to plan benchmark internally first. But right now no guarantee. 15:34:14 <yamahata> mohammed will work on port binding related staff. 15:34:33 <mkolesni> ok 15:35:30 <yamahata> The behaviour of v2 driver would be interesting. If it can be a topic at opendaylight/openstack summit, design summit/forum. 15:36:13 <yamahata> anything else? 15:36:26 <yamahata> I'll create a page, so please add your ideas. 15:36:30 <mkolesni> ok 15:36:55 <yamahata> #topic open mike 15:37:01 <yamahata> anything else to discuss? 15:37:11 <mkolesni> nothing from my side this week 15:37:22 <yamahata> Cool. so we can skip bluejeans 15:37:29 <mkolesni> +1 15:37:31 <vthapar> nothing from me either. 15:37:44 <vthapar> one suggestion, would be good to send openstack voting link to mail list 15:38:00 <yamahata> vthapar: can you take it? 15:38:10 <vthapar> yamahata: sure. 15:38:18 <yamahata> #action vthapar send openstack voting link to mail list 15:38:19 <mkolesni> the list is rather small, perhaps ask people to spread it around 15:38:24 <yamahata> yay ;-) 15:38:39 <vthapar> will send it to neutron, netvirt and ovsdb. 15:38:40 <yamahata> #action everyone ask people to spread openstack voting around 15:38:42 <vthapar> ovsdb is lot bigger :) 15:38:49 <mkolesni> good idea 15:39:19 <vthapar> need to specify which particular topics? the way it is done couldn't extract links to individual submissions. 15:39:31 <vthapar> had to search "Opendaylight" and then find ours. 15:40:33 <yamahata> It seems like that the specific topic can't be linked intentionally. 15:41:07 <vthapar> yeah :) 15:41:28 <mkolesni> perhaps yamahata can send you the names to vote for? 15:41:31 <vthapar> want me to give instructions on how to find, just mention use "opendaylight" in search 15:41:41 <yamahata> mkolesni: sure 15:42:07 <yamahata> The evolution of OpenDaylight with OpenStack: big changes for production environment in future 15:42:16 <yamahata> OpenStack Networking SFC and L2GW service through OpenDaylight 15:42:22 <yamahata> the second is for vthapar 15:42:53 <vthapar> yeah, will add those too. 15:43:26 <vthapar> thsoe are the only two from us, right? 15:43:27 <mkolesni> you can search for "Isaku" ;P 15:43:49 <mkolesni> Have a bite of 1K+ nodes with a handful of servers 15:44:02 <mkolesni> found that one when searching for Isaku :) 15:44:15 <yamahata> there are many odl proposal. SFC seems hot topic 15:44:55 <yamahata> okay anyother topics? 15:45:02 <yamahata> otherwise I'll give you back 15mins. 15:45:06 <mkolesni> +1 15:45:27 <yamahata> thank you everyone! 15:45:33 <mkolesni> thanks Isaku 15:45:35 <yamahata> #topic cookies 15:45:41 <mkolesni> have a nice day everyone 15:45:43 <yamahata> #endmeeting