18:06:05 #startmeeting networking_policy 18:06:06 Meeting started Thu Mar 9 18:06:05 2017 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:06:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:06:09 The meeting name has been set to 'networking_policy' 18:06:27 since igordcard is here, may be we can discuss the QoS patch first 18:06:41 #link QoS via NSPs patch 18:06:44 alright 18:06:59 #topic QoS via NSPs patch 18:07:05 #link https://review.openstack.org/#/c/426436 18:07:45 igordcard: so you mentioned that the NFP tests are failing for an expected reason? 18:08:33 i am seeing some DB errors in the logs 18:08:44 SumitNaiksatam: I think there's still something wrong with the NFP job 18:08:47 #link https://review.openstack.org/#/c/431350/ 18:08:51 presumably from accessing DB objects outside of transactions 18:08:55 that was merged with NFP problems 18:09:01 igordcard: okay, possible 18:09:26 igordcard: if you can confirm that the failures dont have anything to do with your patch, i would be inclined to merge your patch 18:09:30 is songole here? 18:10:31 igordcard: we knew there were going to be transient failures with the NFP job even when we merged the “fix” 18:10:48 igordcard: if its the same issue, then we are good to go ahead with your patch 18:11:32 songole: hi 18:11:36 SumitNaiksatam: the nfp failures included db issues in the link I linked above 18:11:43 hello 18:11:53 igordcard: sorry, linked above as in lined to what? 18:11:57 *linked 18:12:02 #link https://review.openstack.org/#/c/431350/ 18:12:17 igordcard: okay 18:12:30 iirc the GET for the PT is inside a transaction, but I have to look at the code again 18:12:40 I will look again and give more info 18:12:55 igordcard: songole: let me start and offline thread with you, and ash, to get a go ahead from him 18:13:10 ok 18:13:26 igordcard: as mentioned earlier, we need to cut the stable/newton sometime soon, and we want to ideally get this patch in before that 18:13:30 but with the db sync issues in the merged patch above, it might happen that the qos patch is showing a negative consequence of such db sync issues 18:13:41 SumitNaiksatam: yep 18:13:42 igordcard: hmmm 18:13:57 SumitNaiksatam: send me the email and I'll follow through 18:14:09 igordcard: okay thanks, ideally i would like to close this by monday 18:14:15 SumitNaiksatam: unfortunately I really need to go now, this is trip time for me :( 18:14:25 SumitNaiksatam: okay 18:14:29 igordcard: oh sure, thanks for joining 18:14:35 igordcard: enjoy! :-) 18:14:47 :) cya all 18:15:09 #topic Newton sync patches 18:15:11 #link https://review.openstack.org/#/q/topic:newton-sync 18:15:28 I was hoping to come into this meeting saying that we are ready 18:15:58 bad news is that we are not yet, semi good news is that i am down to, what i think, is one remaining issue 18:16:09 SumitNaiksatam: nice! 18:16:14 tbachman: thanks 18:16:31 so if you see this patch form annak: #link https://review.openstack.org/439247 18:16:39 you will see all the integration tests pass 18:16:55 note that the NFP integration tests are a pass through 18:17:32 there is work which needs to be done with the NFP modules before we can get that job to pass, but we are not gating on that for creating the stable/newton branch 18:17:55 the CLI patch (from annak) and heat and UI patches (from me) are also ready 18:18:04 and are being used in the above integration jobs 18:18:23 there are three failures in the py27 job, and they are all related to one issue 18:18:52 i am hoping this is mostly changing the expected values (since the nature and order of notifications has changed, and which is what is being tested) 18:19:23 i promised last week that i would be adding a detailed commit messsage to facilitate the review, unfortunately i couldnt get to it 18:19:25 yet 18:19:56 but once i fix this py27 issue, i will have it in place, and then hopefully i can get some review love from the team 18:20:16 * tbachman nods 18:20:17 annak: i think your db patch and the CLI patches are ready, and can be removed from WIP 18:20:20 tbachman: thanks 18:20:34 SumitNaiksatam: sure, I'll do that 18:21:00 the first patch that will have to be merged obviously is this patch: #link https://review.openstack.org/426437 (once it passes the py27) 18:21:02 annak: thanks 18:21:20 please note that the integration tests will not pass on in my patch #link https://review.openstack.org/426437 18:21:53 they will only pass on anna’s DB patch since it turns out that those changes are necessary in an installation (though not for the UTs to pass) 18:22:25 * tbachman realizes that there are some changes to the APIC and GBP L3 plugins that we probably want to port from upstream neutron 18:22:39 tbachman: okay, lets discuss 18:22:48 tbachman: you mean follow up patches? 18:22:55 yeah — there’s a patch in upstream neutron 18:22:58 let me pull it up 18:23:04 tbachman: okay 18:23:17 this is a revert of a revert: 18:23:17 https://github.com/openstack/neutron/commit/9e7137f7f6c54e19821adf4aad7c3d1fd1e5ad6e#diff-e6ae0f3896f9c7fd927f93cb8bb52174 18:23:53 basically, if you restarted neutron-server, you could get warnings about processes forking after threading 18:23:59 tbachman: okay 18:24:07 they put infra in upstream nuetron to address this 18:24:15 but there are some minor changes to the L3 plugin we should make 18:24:20 tbachman: ah that opens another topic, glad you brought this up 18:24:24 not a biggie, but just something I realized 18:24:29 tbachman: okay 18:24:35 * tbachman can create said patch 18:24:51 so in general, there are things that i have addressed in this patch, and there are things which still need to be addressed 18:25:10 most notably, expect to see a lot of deprecation warnings 18:25:22 i had to also disable some hacking checks 18:25:39 so once we cut the branch, we would need to make these changes as well 18:26:15 and of course, we can work on this parallel, so if anyone wants to jump in and take up some items, please feel free to do so :-) 18:26:39 i will outline, what is missing, in the commit message 18:27:12 thanks in advance for volunteerint :-P 18:27:17 *volunteering 18:27:33 rkukura: any thoughts on what you have seen so far? 18:27:50 tbachman: and rkukura and I have been having offline discussions 18:27:57 had some discussions with annak over emails as well 18:28:07 and some with songole :-) 18:28:34 but here we can all discuss together if we see any issues with how we are moving forward with this 18:28:52 SumitNaiksatam: you mean moving forward with addressing things like the deprecation warnings, etc.? 18:29:04 tbachman: actually i meant what we have so far 18:29:08 ah 18:29:22 I agree on merging the current patches once they are working, and followuping up on other items afterwards 18:29:23 i know that the patches are in WIP 18:29:30 rkukura: okay, thanks 18:29:56 so things were in a flux and not many people would have reviewed 18:30:15 if nothing else on this, we can move on to NFP 18:30:26 sorry, i need to leave now. 18:30:34 songole: oh, okay 18:30:36 I am at a conf 18:30:47 songole: i wanted to discuss the NFP patches with you :-) 18:30:51 We’ll need to decide what kinds of things we continue to backport to mitaka 18:31:05 rkukura: okay 18:31:36 at this point i would imagine only bug fixes 18:32:10 rkukura: you propose being more restrictive even with the bugs? 18:32:25 back-porting actual bug fixes is a given, I think 18:32:30 rkukura: okay 18:33:04 I’m think of ongoing improvements/features/etc., parficularly the APIC AIM stuff and NFP 18:33:11 rkukura: okay 18:33:38 rkukura: i agree that we should try to a model, where stable, means actually stable 18:33:50 *try to move to 18:34:15 SumitNaiksatam: Did you say that we will be a cutting a stable/newton branch right away? 18:34:22 so far we have been driven by user requirements, and unfortunately they have been all over the place in terms of releases 18:34:40 rkukura: after the sync patches merge 18:35:02 Thought so, and newton is considered a long-term release, right? 18:35:27 rkukura: that is right, at least from a RHEL distro perspective 18:35:43 So anything being delivered to customers in the next few months will most likely need to be on master and stable/newton 18:35:56 But maybe not on stable/mitaka 18:36:09 any plans about ocata sync? 18:36:19 rkukura: i dont know about not being on the stable/mitaka, but i would hope not 18:36:29 annak: kicking SumitNaiksatam while he’s barely off the ground ;) 18:36:30 lol 18:36:44 :)) just curious :) 18:36:54 :-) 18:37:00 i was hoping no one would bring that up 18:37:03 just kidding 18:37:04 lol 18:37:34 do we have any idea how much additional work would be involved in supporting ocata? 18:38:04 rkukura: unfortunately i dont, and given the current experience i dont want to make any guesses :-( 18:38:24 hopefullly less that ocata was a shorter cycle 18:38:32 annak: we can start ocata sync immediately after we cut stable/newton 18:38:40 rkukura: finger crossed 18:38:53 ok 18:39:23 annak: to give you a historical perspective, we cut stable/mitaka which neutron newton was still worked on 18:39:50 annak: and my hope at that time was to immediately get GBP to sync with newton even before neutron stable/newton was cut 18:39:58 and then fast forward to today :-) 18:40:22 * tbachman thanks SumitNaiksatam for taking on all the pain that is syncing with upstream 18:40:33 tbachman: thanks, but it was team effort 18:40:47 during that time we focussed on NFP, and the drivers (mostly AIM) 18:41:04 anyway, we will skip the NFP discussion since songole is not here 18:41:13 there are a bunch of pending patches 18:41:21 #topic Open Discussion 18:41:25 anything else for today? 18:41:30 :) its a huge project and very few people.. 18:41:31 SumitNaiksatam: it appears a bunch have just merged 18:41:50 annak: we just picked up one more :) 18:41:57 annak: you are already making a big contribution :-) 18:42:21 hopefully I'll be able to continue 18:42:37 annak: definitely looking forward to it 18:42:52 thanks :) me too 18:42:58 alrighty if nothing else, we can wrap up, and get back to doing real work :-) 18:43:02 not sure of the timing on this, but are there any potential GBP-related sessions at the Boston summit? 18:43:13 rkukura: good point 18:43:27 rkukura: i think there was some talk about doing something on NFP, would need to check 18:43:42 but, in general yeah, good time to plan 18:44:23 annak: you can consider a presentation as well 18:44:35 based on the work you have done 18:44:58 still not sure I'll make it there.. 18:44:58 or a a joint presentation 18:45:06 rkukura: right 18:45:11 annak: okay, we can discuss offline 18:45:23 ok 18:45:28 alirghty, thanks all for joining 18:45:33 SumitNaiksatam: thanks! 18:45:35 see you next week, bye! 18:45:40 #endmeeting