15:08:28 <manjeets> #startmeeting opendaylight_neutron 15:08:28 <odl_meetbot> Meeting started Mon Jul 16 15:08:28 2018 UTC. The chair is manjeets. Information about MeetBot at http://ci.openstack.org/meetbot.html. 15:08:28 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:08:28 <odl_meetbot> The meeting name has been set to 'opendaylight_neutron' 15:09:15 <manjeets> #topic Agenda 15:09:21 <manjeets> what do we have today, 15:09:23 <manjeets> ? 15:09:39 <manjeets> CI, SBI consistency and L3 flavors ? 15:09:45 <mpeterson> also tempest 15:09:55 <mpeterson> and fluorine tests 15:09:59 <mkolesni_> tempest narrowing 15:10:05 <mkolesni_> fullstack 15:10:11 <manjeets> #topic CI 15:10:58 <mpeterson> okey, so if no one is starting I will 15:11:05 <mpeterson> two things with it 15:11:13 <manjeets> I almost got stable/queens CI fixed last week, there was only failure left 15:11:15 <mpeterson> 1) the fullstack tests were disabled 15:11:32 <mpeterson> 2) stable/* are broken 15:11:51 <mpeterson> manjeets: cool 15:12:08 <manjeets> mpeterson, this https://review.openstack.org/#/c/579915/ will stable queens I need to fix requirements jobs 15:12:31 <manjeets> before last run it was all passing except requirements 15:13:07 <mpeterson> manjeets: let's recheck it 15:13:21 <mpeterson> manjeets: there was a CI general error 15:13:31 <mpeterson> manjeets: recheck sent 15:13:37 <manjeets> thanks mpeterson 15:13:51 * manjeets hand mic to mpeterson to continue other CI stuff 15:14:14 <mpeterson> so for (1) is just informative, and when we move to that topic mkolesni_ can expand 15:14:34 <mpeterson> (2) we'll need someone to fix it but right now it's not a priority nor a requirement 15:14:50 <mpeterson> so if anyone wants to take it upon themselves, be my guest 15:14:55 <mkolesni_> mpeterson, manjeets is fixing queens 15:15:10 <mpeterson> mkolesni_: yeah, I figured :) 15:15:41 <mkolesni_> as for fullstack 15:15:43 <mkolesni_> it broke 15:15:47 <mkolesni_> but i dont know why 15:15:59 <mkolesni_> anyway i dont see any value in these tests 15:16:25 <mkolesni_> theyre too white box, and mimic ODL behavior that can be different 15:16:31 <manjeets> mkolesni_, they're just bunch of tests inherited from neutron ? 15:16:38 <mkolesni_> so i disabled them 15:16:48 <mkolesni_> no i think you and/or isaku wrote them? 15:17:06 <manjeets> mkolesni_, It wasn't me 15:17:29 <mkolesni_> ok 15:17:32 <mpeterson> I'm super okey with just removing them altogether 15:17:36 <mkolesni_> anyway theyre broken now 15:17:46 <mkolesni_> so i suggest to delete them as they add no value 15:17:55 <manjeets> mkolesni_, If I recall at that time it was done for stadium requirement 15:18:01 <mkolesni_> i have to go for a coyuple of minutes, please discuss and ill be back promptly 15:18:46 <mpeterson> this is what the requirement says: exhaustive OpenStack CI coverage (unit, functional, tempest), using OpenStack CI (upstream) resources (this means using infra CI, not just downstream CI, as Grafana support will be required); 15:19:36 <mpeterson> manjeets: from that I gather that we are okey with just leaving the functional and deleting fullstack which IMHO is just an useless invention 15:20:11 <mkolesni_> yes afaik theres no such requirement 15:20:17 <manjeets> mpeterson, before I can say anything, I'd like to read code at least once before I give my opinion, so can we wait until next week for removal ? 15:20:21 <mpeterson> manjeets: actually, if we talk about stadium requirements, currently the most pressing is getting granade to work 15:20:28 <mpeterson> manjeets: I'm super okey with that 15:20:34 <mkolesni_> fullstack may fit neutron native drivers where the implementation is known 15:20:44 <mkolesni_> manjeets, sure 15:21:15 <mkolesni_> manjeets, anyway for now theyre non voting to unbreak the ci 15:21:42 <mkolesni_> shall we move on? 15:21:52 <manjeets> ok, making them non voting would give me some time 15:21:54 <manjeets> sure 15:21:57 <mpeterson> yeah, the only thing is what I just say and we can make it #3\ 15:22:09 <mpeterson> 3) we need to fix grenade as it is part of the stadium reqs 15:22:13 <mkolesni_> mpeterson, ok 15:22:16 <mkolesni_> tempest narrowing - https://review.openstack.org/582176 15:22:26 <mkolesni_> so i did some thinking with mpeterson last week 15:22:31 <mkolesni_> and then some investigation 15:22:47 <mkolesni_> after a battle to get devstack to work :) 15:23:02 <manjeets> mpeterson, I fixed it several times with hacks but not kept up, If someone wants to take over I'd be happy to help 15:23:10 <mkolesni_> i tested what happens to tempest tests when the ODL is down 15:23:11 <manjeets> grenade I mean in this ^^ 15:23:29 <mpeterson> manjeets: okey, I'll see if I have bw, for now let's say I'm assigned to it 15:23:37 <mkolesni_> so turns out only a few dozens are failing and almost all from the scenario suites 15:24:15 <mkolesni_> so in order to have a more stable ci, and less interference from api tests (we dont even touch the api in any way) 15:24:30 <mkolesni_> i have this patch to narrow down the tests to just the scenario tests 15:25:05 <manjeets> mkolesni_, so volume api tests are disabled ? 15:25:07 <mkolesni_> and api.compute which have quite a few cases that bring up a vm so they're also relevant 15:25:21 <mkolesni_> yes that's still disabled by the blacklist anyway 15:25:43 <mkolesni_> http://git.openstack.org/cgit/openstack/networking-odl/plain/tempest-blacklist.txt 15:25:58 <mkolesni_> but now only the more interesting tests will run 15:26:18 <mkolesni_> it's still circa 300 tests, but at least 700 irrelevant ones wont run 15:26:47 <mkolesni_> i feel this is a good compromise, but we can further investigate and narrow down the list some more 15:27:00 <manjeets> mkolesni_, it will save us time and some life ... lol 15:27:07 <mpeterson> and on the plus side, it seems that now tempest will be stable and thus we can enable it 15:27:07 <mkolesni_> yes also :) 15:27:15 <mkolesni_> run time seems to reduce by ~30 mins 15:27:25 <mkolesni_> well, more stable :) 15:28:03 <manjeets> cool thanks mkolesni_ for working on it 15:28:19 <mkolesni_> sure np i just want to have a good env :) 15:28:39 <mkolesni_> so if it looks good lets merge and next we can reenable the stable of the tempest tests 15:29:28 <mkolesni_> also the blacklist was trimmed but if i missed something obviousl let me know 15:29:41 <manjeets> mkolesni_, its very important to run useful ones and keep it stable rather than running every single scenario and burn ourselves 15:30:00 <mkolesni_> yes 15:30:21 <mkolesni_> generally lets start with this and if we see unuseful ones lets trim thre fat 15:30:32 <mpeterson> +1 15:30:34 <manjeets> +1 15:30:40 <mpeterson> what's next in the agenda? 15:30:53 <manjeets> SBI consistency 15:30:55 <mpeterson> oh, related to this 15:30:55 <mpeterson> wait 15:30:57 <mpeterson> don't change 15:31:04 <mkolesni_> ? 15:31:12 <mpeterson> there are 2 more things related to CI that I want to introduce 15:31:17 <manjeets> sure 15:31:27 <mkolesni_> please 15:31:29 <mpeterson> 5) can we reenable rally? why was that disabled? 15:32:18 <mkolesni_> actually u know rally doesnt test odl is working right? we only should have there tests that launch a vm and ping it 15:32:26 <mkolesni_> i think theres a lot of fat there also 15:32:50 <mkolesni_> i can investigate this on my env and push a patch to reduce to the interesting tests 15:33:39 <mkolesni_> then we can make them voting again :) 15:34:07 <mpeterson> mkolesni_: sounds good to me 15:34:16 <mkolesni_> manjeets, ? 15:34:23 <manjeets> sounds good to me as well 15:34:28 <mkolesni_> cool 15:34:31 <mkolesni_> ill get on it then 15:34:34 <mpeterson> 6) can we enable the fluorine tests? they were introduced but never made voting 15:34:45 <mkolesni_> #action mkolesni_ will trim down rally and renable it 15:35:01 <manjeets> is the pass rate good enough to be voting ? 15:35:02 <mkolesni_> mpeterson, how stable are they? 15:37:07 <mpeterson> mkolesni_: they are stable IMHO 15:37:33 <mkolesni_> mpeterson, check grafana and if it seems stable make them voting 15:37:58 <manjeets> #action mpeterson to check fluorine and make them voting 15:38:12 <mpeterson> I've just checked, they are stable 15:38:16 <mpeterson> I'll commit a patch 15:38:19 <mkolesni_> cool 15:38:30 <mkolesni_> lets also discuss l3 flavor 15:38:41 <mpeterson> okey, we can move to either SBI or L3 15:38:50 <manjeets> #topic L3_flavors 15:39:20 <manjeets> I addressed most of the comments over the patch, I guess i need to address couple more and it will be close to go then ? 15:39:25 <manjeets> mkolesni_, ^ 15:39:50 <mkolesni_> manjeets, u missed these left at PS35 15:39:50 <mkolesni_> PS29, Line 128: I don't see it in PS#35, did you miss to submit it maybe? 15:39:51 <mkolesni_> PS29, Line 217: I don't see it in PS#35, did you miss to submit it maybe? 15:40:24 <mkolesni_> also mpeterson had a few 15:40:34 <mkolesni_> but i think its close to merge 15:40:40 <manjeets> mkolesni_, I'll read them and address it today 15:40:45 <mkolesni_> cool 15:40:52 <manjeets> #action manjeets to address comments on l3 flavors 15:40:57 <manjeets> lets move on 15:41:03 <manjeets> #topic SBI consistency 15:41:32 <manjeets> sorry, I must admit I didn't get bw to review it last week 15:41:40 <mpeterson> okey, so re that, enginefacade was merged so that's good 15:41:44 <mkolesni_> im reviewing it as much as i can, would like your opinion also 15:41:50 <mpeterson> now we need to put as much effort as possible on SBI too 15:42:07 <manjeets> mpeterson, yes 15:42:09 <mpeterson> the first and perhaps the most important patch has been already +2 by mkolesni_ 15:42:44 <mkolesni_> indeed it looks good to me 15:43:29 <mpeterson> and with tempest narrowed down that will help a lot with the v3 driver 15:43:29 <manjeets> mkolesni_, and mpeterson I'd like to bring one thing in to your notice 15:43:33 <manjeets> https://review.openstack.org/#/c/578225/ 15:43:47 <mpeterson> so SBI (all commited AFAIC) and v3 are my priorities now 15:43:49 <manjeets> would this make sense for n-odl ? 15:43:59 * mpeterson is looking at the review 15:44:20 <manjeets> it claims to remove the need to sync db 15:45:36 <mkolesni_> what need exactly? the polling/websocket? 15:46:43 <manjeets> mkolesni_, I haven't investigated it much, but it was just FYI, we can discuss it later sometime 15:47:53 <mpeterson> manjeets: I'm not sure I understand the need either... let's review the concept and talk this again next week 15:48:40 <manjeets> ok lets move on 15:49:09 <manjeets> i think we discussed all the topics 15:49:15 <manjeets> #topic open_mic 15:49:21 <mkolesni_> i have nothing else to add 15:49:28 <mkolesni_> we have a big plate :) 15:49:38 <mkolesni_> btw feature freeze is next week right? 15:50:03 <manjeets> mkolesni_, i think so rocky-3 is week away 15:50:18 <mkolesni_> also do u have confirmation re PTG yet? 15:50:55 <mpeterson> nothing to add either :) and +1 to mkolesni_ 's question 15:51:04 <manjeets> mkolesni_, my manager said he's working on it, most likely i'll go but you never know until final confirmation 15:51:17 <mkolesni_> of course 15:51:21 <mkolesni_> i have no news either 15:51:23 <manjeets> mpeterson, mkolesni_ you guys ? 15:51:25 <manjeets> ok 15:51:27 <mkolesni_> will update if i do 15:51:46 <mpeterson> same status 15:52:03 <manjeets> cool thanks then have good rest of the week :) 15:52:08 <mkolesni_> thank you too 15:52:13 <mkolesni_> cya guys! 15:52:20 <manjeets> #endmeeting