15:01:27 <ihrachys> #startmeeting neutron_upgrades
15:01:28 <openstack> Meeting started Mon Mar 27 15:01:27 2017 UTC and is due to finish in 60 minutes.  The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:31 <ihrachys> o/
15:01:31 <manjeets_> o/
15:01:32 <openstack> The meeting name has been set to 'neutron_upgrades'
15:01:37 <electrocucaracha> o/
15:02:19 <sshank> o/
15:02:20 <ihrachys> #link https://wiki.openstack.org/wiki/Meetings/Neutron-Upgrades-Subteam Agenda
15:02:35 <ihrachys> #topic Announcements
15:03:01 <ihrachys> nothing specific to upgrades, just a note that Pike-1 is in 2-3 weeks. it doesn't change much in what we do and track, so moving on.
15:03:08 <ihrachys> #topic Linuxbridge multinode grenade job
15:03:12 <ihrachys> manjeets_: your stage
15:03:21 <manjeets_> ihrachys, tried couple of things there
15:03:32 <manjeets_> forcing concurrency = 2
15:03:41 <manjeets_> seems like that is not the issue
15:04:04 <manjeets_> I've increased the test timeout but unfortunately did not get job placed on rax node yet
15:04:38 <manjeets_> i've checked timestamps it is taking so much time to bring up fip and test timesout
15:05:04 <ihrachys> ok, let's say you indeed see that it shows up on rax only. what next? you can't just stop landing the job there, right?
15:05:24 <manjeets_> it fails only on rax-node cloud, I've made one other patch dependent on patch where I exported increased time out
15:05:29 <ihrachys> manjeets_: how do you know that it indeed takes long to set up fip? do you see router updates slowly handled?
15:05:50 <manjeets_> ihrachys, I checked timestamps from tests
15:05:54 <ihrachys> I mean in agent logs
15:06:05 <manjeets_> and compared when it succeed vs fail
15:06:12 <ihrachys> well tempest output won't show you if FIP is configured by the agent or now
15:06:14 <ihrachys> *or not
15:06:52 <ihrachys> to know that, you may need to find exact place in the agent log where the router update event corresponding to FIP creation/attachment is handled in agent side
15:07:58 <manjeets_> ohk I'll check l3 agents logs today
15:08:31 <ihrachys> cool, let me know if you struggle finding where the router update event handler is on the agent side
15:09:12 <manjeets_> ok
15:10:23 <ihrachys> #topic Mixed server versions
15:10:40 <ihrachys> I had an action item to follow up here, writing the spec. it's still not in gerrit. :-x
15:10:57 * ihrachys was consumed by other internal matters for most of the week
15:11:56 <ihrachys> also dolphm in the RFE expressed some questions about how critical the proposal is: https://bugs.launchpad.net/neutron/+bug/1672852
15:11:56 <openstack> Launchpad bug 1672852 in neutron "[RFE] Make controllers with different list of supported API extensions to behave identically" [Wishlist,New] - Assigned to Ihar Hrachyshka (ihar-hrachyshka)
15:11:57 <electrocucaracha> but at least it's only matters of provide documentation for deployers right?
15:12:08 <ihrachys> suggesting affinity aware LBs used in all production envs
15:12:37 <ihrachys> electrocucaracha: no I am talking about the RFE ^
15:12:59 <ihrachys> something internal to neutron-server that makes all cluster nodes to report identical /extensions/
15:13:35 <ihrachys> dolphm: also has concerns about the complexity of this thing. I guess we will need to resolve those in gerrit when the spec is up.
15:13:44 <manjeets_> afaik spec or rfe for maintaining minimum subset of extensions b/w mixed versions
15:13:52 <manjeets_> right ?
15:14:07 <ihrachys> yes
15:16:11 <ihrachys> there is also somewhat related online data migration CLI command: https://review.openstack.org/#/c/432494/
15:16:22 <ihrachys> electrocucaracha: I reviewed it like 15 mins before the meeting :)
15:16:48 <dolphm> ihrachys: just caught up on your comments in that bug; i just wanted to voice alternative approaches and see the pro's and con's weighed out, not push you toward one solution. it sounds like your approach is pretty welll thought out as-is
15:17:00 <ihrachys> electrocucaracha: there is definitely some code polishing to do in the implementation, also the mechanism of registering new migrations is not ideal, we will need to open it for external plugins.
15:17:29 <ihrachys> dolphm: I will try to capture your comment in the spec and will reach out to you for feedback.
15:18:07 <dolphm> ihrachys: i'm not sure i've seen the spec; link?
15:18:14 <dolphm> (it's not linked in the bug)
15:18:15 <ihrachys> dolphm: there is none just yet
15:18:19 <dolphm> ihrachys: gotcha
15:18:20 <ihrachys> it's in progress
15:18:25 <dolphm> ihrachys: yeah, ping me when it's up for review
15:18:51 <electrocucaracha> ihrachys: ok, I'm going to check your comments and thanks dolphm for the feedback on that patch
15:18:52 <ihrachys> electrocucaracha: what I basically wanted for the plugins is so that we don't need to modify the tool when an external plugin wants to define their own migration
15:19:24 <ihrachys> ack, thanks for working on it
15:19:34 <ihrachys> #topic Object implementation
15:19:45 <ihrachys> https://review.openstack.org/#/q/topic:bp/adopt-oslo-versioned-objects-for-db
15:19:51 <ihrachys> quotas landed: https://review.openstack.org/338625
15:20:19 <ihrachys> also gate fix for unit tests landed: https://review.openstack.org/#/c/447600/
15:21:45 <ihrachys> there is also LIKE patch that we wanted to get back but I failed: https://review.openstack.org/#/c/419152/
15:21:55 <ihrachys> I see electrocucaracha left some comments, nice
15:23:36 <electrocucaracha> ihrachys: I haven't had time to review or create an UT for unique keys test data generation
15:25:04 <electrocucaracha> and in terms of metrics we have 82% of the model classes have their corresponding OVO class merged in master
15:25:30 <electrocucaracha> and 53% are using OVO to do CRUD operations
15:25:54 * electrocucaracha long journey
15:26:41 <ihrachys> well not that bad, I thought it's lower ;)
15:26:56 <ihrachys> gotta get more serious about NetworkSegment, SG and such
15:26:58 * manjeets_ thought it must be around 65 for usage
15:27:48 <electrocucaracha> ihrachys: well, actually Allocation and Endponits will give us more
15:27:58 <electrocucaracha> https://review.openstack.org/#/c/367810/
15:29:03 <ihrachys> it's WIP
15:29:13 <ihrachys> dasanind: are you working on it?
15:29:34 <dasanind> ihrachys: Yeah I am working on it
15:30:05 <dasanind> ihrachys: I put a WIP for the lock mode. Will respin it this week and complete it
15:30:07 <ihrachys> I see it's still using locsk
15:30:11 <ihrachys> yea
15:30:20 <ihrachys> ok, send it my way when ready
15:30:26 <dasanind> ihrachys: ack
15:31:52 <ihrachys> #topic Other patches on review
15:31:58 <ihrachys> https://review.openstack.org/#/q/status:open+message:%22UpgradeImpact%22+project:openstack/neutron
15:32:11 <ihrachys> the haproxy switch gone in, so no longer in the query
15:32:26 <ihrachys> there is nothing else to discuss here really, let's move on
15:32:36 <ihrachys> #topic Review priorities for the week
15:32:49 <ihrachys> electrocucaracha: I figured it helps to have that discussion weekly
15:33:39 <manjeets_> review priorities discussion ?
15:33:41 <electrocucaracha> ihrachys: definitely,
15:33:44 <ihrachys> ofc, the spec for /extensions handling for mixed server versions and the CLI are in the focus: https://review.openstack.org/#/c/432494/
15:33:46 <ihrachys> manjeets_: yes
15:33:49 <manjeets_> ++
15:33:55 <electrocucaracha> ihrachys: I was checking the spreadsheet https://docs.google.com/spreadsheets/d/1FeeQlQITsZSj_wpOXiLbS36dirb_arX0XEWBdFVPMB8/edit#gid=1434170112
15:34:24 <ihrachys> beyond, that, realistically, we should be able to land the LIKE support: https://review.openstack.org/419152 and maybe endpoints if dasanind is quick to respin: https://review.openstack.org/367810
15:34:40 <ihrachys> all of those are still work in progress
15:34:46 <ihrachys> anything ready to land NOW?
15:35:13 <sshank> Router Extra Attrs integration seems like its ready.
15:35:28 <sshank> Is it in priority?
15:35:41 <electrocucaracha> I remember that Meterings was fine, but I need to address the dasm comments
15:36:10 <ihrachys> sshank: it's not priority but if it's ready it can get focused attention
15:36:20 <ihrachys> the link to extra_attrs patch: https://review.openstack.org/#/c/381209/
15:36:33 <ihrachys> electrocucaracha: https://review.openstack.org/#/c/350387/? it's abandoned
15:37:02 <dasm> ihrachys: i think electrocucaracha mentioned about this one: https://review.openstack.org/#/c/433125/
15:37:24 <ihrachys> oh ok
15:37:29 <electrocucaracha> ihrachys: for no reason he abandoned but I reopened a new one ^^
15:37:30 <ihrachys> what's the relation to the former?
15:37:40 <ihrachys> ok we will need to maintain the authorship
15:37:45 <ihrachys> Co-Authored-By
15:38:02 <electrocucaracha> ihrachys: well, he decided to abandoned because he was not doing anything
15:38:29 <electrocucaracha> ihrachys: I took that patch since the beginning
15:38:40 <electrocucaracha> but I can add his name
15:38:45 <ihrachys> electrocucaracha: sure. all I am saying is, if you reused the code, make sure the author is mentioned in the commit message
15:39:03 <electrocucaracha> ihrachys: ok, no problem
15:40:02 <ihrachys> ok I think realistically it's already more than enough to review for the week
15:40:37 <ihrachys> #topic Open discussion
15:41:09 <ihrachys> anything specific to discuss that was not covered in any other section?
15:41:37 <sshank> ihrachys, I think the error similar to session expunge in port binding level is seen in router l3 binding integration as well.
15:41:40 <electrocucaracha> talking about dates, when starts the next release?
15:42:21 <ihrachys> electrocucaracha: you mean Queens? it's like +6 months since PTG
15:42:29 <ihrachys> which was a month ago
15:43:12 <ihrachys> sshank: ack
15:44:17 <electrocucaracha> ihrachys: ok, in that case we still have time for landing more patches, as long as we keep the same cadence
15:45:02 <manjeets_> electrocucaracha, ++
15:45:21 <ihrachys> +
15:45:44 <ihrachys> ok let's use the 15 mins for something useful like drinking coffee ;)
15:45:51 <ihrachys> thanks everyone
15:45:57 <electrocucaracha> thanks ihrachys
15:46:01 <ihrachys> #endmeeting