14:00:43 <ihrachys> #startmeeting neutron_upgrades
14:00:47 <openstack> Meeting started Thu Oct  5 14:00:43 2017 UTC and is due to finish in 60 minutes.  The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:48 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:50 <annp> hi
14:00:50 <ihrachys> hey everyone
14:00:51 <openstack> The meeting name has been set to 'neutron_upgrades'
14:00:57 <TuanVu> hi everybody
14:01:00 <lujinluo> o/
14:01:26 <ihrachys> we had gate broken (and still in bad shape) for the last week
14:01:31 <ihrachys> because of zuul v3 migration
14:01:50 <ihrachys> so reviews/merges are expected to be slow
14:02:03 <ihrachys> but infra steadily works on existing issues
14:02:26 <ihrachys> I'm saying that mostly to manage expectations :)
14:02:37 <lujinluo> ;)
14:02:37 <ihrachys> anyway, let's dive in patches
14:02:45 <ihrachys> https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/adopt-oslo-versioned-objects-for-db
14:02:59 <ihrachys> slaweq updated https://review.openstack.org/#/c/411981/
14:03:23 <ihrachys> and I +2; we will wait a bit for other reviewers; if they don't show up, I will fish for reviews.
14:03:47 <ihrachys> next is https://review.openstack.org/#/c/495810/ "Use Agent OVO in agents_db and test_agents_db"
14:04:23 <TuanVu> I'm waiting for response from Slawek
14:05:28 <ihrachys> TuanVu, I think he means that some plugins may use the method directly
14:05:36 <ihrachys> and may assume result is models not OVO
14:05:40 <ihrachys> see here: http://codesearch.openstack.org/?q=get_agents_db&i=nope&files=&repos=
14:06:58 <ihrachys> it may be that those cases don't really care about whether it's model or OVO though
14:06:59 <annp> So we shouldn't change it at lease this time. right?
14:07:04 <ihrachys> at least on first sight that's the case
14:07:37 <TuanVu> ok, thank you Ihar
14:07:40 <TuanVu> I got it
14:07:43 <annp> ok.
14:07:55 <ihrachys> annp, I think it's just worth looking at the methods where it's used (in those two other repos) and see if it still makes sense with OVO
14:08:43 <annp> ihrachys:  It make sense. Thanks.
14:08:43 <ihrachys> f.e. if they would set an attribute on the resulting models and assume that changes would flush on transaction context manager exit, then it would be a problem
14:08:55 <ihrachys> but if they e.g. just extract ['id'] then it should work for both
14:10:06 <annp> yes. please go ahead.
14:10:11 <ihrachys> ok
14:10:58 <ihrachys> next is https://review.openstack.org/#/c/382037/ "WIP Integration of Port Binding Level OVO"
14:11:14 <ihrachys> it's wip
14:11:23 <ihrachys> I see slaweq rebased it and will look at results
14:11:36 <ihrachys> moving on
14:11:42 <ihrachys> https://review.openstack.org/#/c/501925/ "Use Agent OVO in l3_agentschedulers_db"
14:12:14 <ihrachys> I see it's not based on latest patch for agents_db
14:12:18 <ihrachys> should we rebase
14:12:19 <ihrachys> ?
14:13:31 <annp> yes, it should be rebase. tomorrow, i will rebase it.
14:13:55 <ihrachys> ok
14:14:11 <ihrachys> also to the question for TODO https://review.openstack.org/#/c/501925/3/neutron/db/l3_agentschedulers_db.py@400
14:14:32 <ihrachys> I don't think it would significantly help if we go with that other patch
14:15:03 <annp> ok. i will remove it also.
14:15:11 <annp> :)
14:15:14 <ihrachys> that's from code simplicity perspective
14:15:28 <ihrachys> but I wonder, isn't configurations supposed to be a dict?
14:15:46 <ihrachys> so matching it against a string-contains is a bit weird
14:17:30 <ihrachys> what would be nice to have is a method that would construct that "x: y" string from kwargs
14:17:36 <ihrachys> but that's probably out of scope for the patch
14:18:50 <annp> actually, I didn't think about that
14:19:13 <annp> So let us dig more and comment in gerrit.
14:19:17 <annp> Is it ok?
14:19:22 <ihrachys> ok
14:19:26 <ihrachys> I added it to my queue
14:19:34 <ihrachys> next patch is https://review.openstack.org/#/c/321001/ "objects: get, update and delete converted to Subnet OVO usage"
14:19:46 <ihrachys> slaweq respinned it and it passed the gate
14:19:51 <ihrachys> which is kinda incredible
14:19:59 <ihrachys> he was not sure whether he did everything right
14:20:35 <lujinluo> lol
14:20:44 <ihrachys> I will need to have a look, that's all I have to say on that so far
14:21:33 <ihrachys> next is https://review.openstack.org/#/c/501190/ "Integration of L3HARouterAgentPortBinding in ml2/drivers/l2pop/db.py"
14:22:45 <ihrachys> I see TuanVu replied there to my suggestion saying there is some error
14:22:53 <ihrachys> in general, we should avoid passing sessions around
14:23:03 <ihrachys> so we
14:23:13 <ihrachys> so we'll need to look closer at error you get
14:23:19 <ihrachys> I don't think revert to patchset 8 is a good idea
14:23:52 <ihrachys> TuanVu, the error you get - http://paste.openstack.org/show/622102/ - suggest you pass wrong context into the function
14:24:01 <ihrachys> PortContext != NeutronContext
14:24:03 <TuanVu> thanks, it would be great if we could have your comment on that error
14:24:35 <ihrachys> ok sure
14:25:08 <ihrachys> TuanVu, can you upload the latest version you have that triggers the error?
14:25:30 <ihrachys> so that we can discuss more specifically
14:25:34 <TuanVu> ok, sure
14:25:42 <ihrachys> great
14:25:42 <TuanVu> I'll upload it tomorrow
14:25:52 <TuanVu> thank you in advance
14:26:24 <ihrachys> next are two network ovo patches that I will skip since they are wip
14:26:30 <ihrachys> unless you have some specific questions
14:27:06 <ihrachys> ok moving to the next
14:27:13 <ihrachys> https://review.openstack.org/#/c/396351/ "Integration of Floating IP OVO"
14:27:28 <ihrachys> there's a bunch of comments there that need to be addressed
14:27:37 <lujinluo> i have a local patch now. should be able to push it next week..
14:27:45 <ihrachys> geat
14:27:47 <ihrachys> *great
14:27:50 <lujinluo> occupied by some internal work this week
14:27:57 <lujinluo> sorry for that
14:28:13 <ihrachys> np, that's what feeds us all :)
14:28:52 <ihrachys> next patch that seems ready is https://review.openstack.org/#/c/466237/ "Change join relationship between routerport and router"
14:29:12 <ihrachys> seems green, just needs reviews it seems
14:29:55 <ihrachys> and that's it for patches that are not wip/in bad shape
14:30:14 <ihrachys> there is plenty of review work for me there, thanks for keeping it coming :)
14:30:26 <TuanVu> thank you Ihar :)
14:30:32 <lujinluo> thank you for all the insightful reviews!
14:32:04 <ihrachys> np
14:32:36 <ihrachys> on gate side, we seem to have dvr-multinode grenade non-voting because of the dvr fiasco of Pike release time
14:32:47 <ihrachys> I had a revert here https://review.openstack.org/#/c/505318/ now that we fixed dvr issues
14:33:02 <ihrachys> but now that we moved to zuulv3 I will need to apply the same to its jobs instead
14:33:51 <ihrachys> infra added a lot of work on dev's plates with the move :)
14:34:05 <ihrachys> but it should be for the better in the end
14:34:15 <ihrachys> ok, I have nothing else to discuss. do you?
14:34:32 <lujinluo> none from me
14:34:39 <TuanVu> me either
14:35:08 <ihrachys> ok great. thanks everyone for all the patches, it's enormous to see it moving forward in such a pace.
14:35:16 <ihrachys> have a nice evening
14:35:19 <ihrachys> #endmeeting