15:02:05 <ihrachys> #startmeeting neutron_upgrades
15:02:06 <openstack> Meeting started Mon Feb 29 15:02:05 2016 UTC and is due to finish in 60 minutes.  The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:02:10 <openstack> The meeting name has been set to 'neutron_upgrades'
15:02:17 <korzen> hello \o
15:02:17 <ihrachys> hi neutrinos ;)
15:02:21 <rossella_s> hi!!
15:02:24 <mhickey> Hello
15:02:24 <sayalilunkad> hello
15:02:46 <ihrachys> #topic Code Sprint (Mar 14-16, Brno)
15:03:06 <ihrachys> #link https://etherpad.openstack.org/p/code-sprint-neutron-objects-brno
15:03:27 <ihrachys> I see some folks signed up there already
15:03:31 <ihrachys> some are tentative
15:03:34 <electrocucaracha> hey
15:03:40 <sayalilunkad> is there public transport from the airport to the venue?
15:03:42 <ihrachys> mhickey: are you up to get to Brno? :)
15:03:55 <mhickey> ihrachys: iS bRNO RELATIVELY SMALL. sO WOULD ANY HOTEL IN THE CITY DO?
15:03:57 <ihrachys> sayalilunkad: the etherpad, starting at line 7
15:04:21 <mhickey> ihrachys: Travel approved. Need to book it now! :)
15:04:23 <ihrachys> mhickey: it's quite small, but I suggest you get Continental, that is in line 18 of the etherpad
15:04:29 <ihrachys> the price is more reasonable, and location is ok.
15:04:32 <ihrachys> mhickey: great!
15:04:47 <ihrachys> sayalilunkad: I see you on the list. have you booked it already?
15:04:51 <mhickey> ihrachys: I will try but it would have to be approved internally.
15:04:52 <sayalilunkad> ihrachys: yes
15:05:09 <ihrachys> sayalilunkad: great :) see you here in two weeks.
15:05:17 <sayalilunkad> yup :)
15:05:35 <korzen> I have room booked in Continental, good price and we can commute together, taxi or bus
15:05:52 <rossella_s> I am also staying at the continental
15:06:01 <sayalilunkad> that would be nice
15:06:05 <ihrachys> note to everyone: we have some better price option for one of the hotels, details in the etherpad starting at line 21. I would like to ask everyone who wants to use the option to book it this week. next week I will need to close the offer.
15:07:05 <rossella_s> ihrachys, thanks for taking care of this! the price is very good indeed
15:07:05 <ihrachys> overall, if you have some questions on the sprint that are not covered in the etherpad, please don't hesitate to ask me in private, I will try to accommodate
15:07:16 <mhickey> ihrachys: sorry for caps mess ^^^^
15:07:23 <ihrachys> my pleasure
15:07:24 <ihrachys> ok great, now to more juicy matters ;)
15:07:46 <ihrachys> #topic Partial Multinode Grenade
15:08:03 <ihrachys> the job is now in neutron check queue woohoo! :)
15:08:06 <ihrachys> still not voting
15:08:18 <ihrachys> we will need to wait a bit to make sure it's stable to enable votes
15:08:24 <ihrachys> but seems good so far
15:08:32 <ihrachys> #link http://grafana.openstack.org/dashboard/db/neutron-failure-rate
15:08:44 <ihrachys> the link should show the rate of failures comparing to other jobs
15:09:04 <korzen> I'm still not having the +2 for multinode DVR job, need to advertise a little bit more
15:09:16 <korzen> #link https://review.openstack.org/#/c/250215
15:09:28 <ihrachys> korzen: yeah, infra stuff takes a while sometimes
15:09:41 <ihrachys> I guess you pulled over folks that could help us, like Anita or Sean?
15:10:03 <korzen> yes, I've ping them a short while ago
15:10:20 <saisriki> hello all
15:10:22 <korzen> I guess they have other more important issues
15:10:23 <ihrachys> ok, let's wait on it a bit more and then try to ping again
15:10:24 <ihrachys> saisriki: hi
15:10:53 <ihrachys> korzen: yeah, people are dragged by release stuff right now
15:11:23 <ihrachys> ok, apart from that, I guess we just wait for that dvr flavour job in experimental queue to proceed on it
15:11:24 <ihrachys> #topic Object implementation
15:11:42 <ihrachys> I'd like to remind to everyone we use 'ovo' tag
15:11:45 <ihrachys> #link https://review.openstack.org/#/q/topic:ovo
15:12:02 <ihrachys> if you'd like your patches to get more attention from folks, please make sure the topic is correct
15:12:38 <ihrachys> also note that we are approaching Mitaka release, so we try to refrain from touching any code that is not objects specific
15:13:06 <ihrachys> so we split patches that introduce new objects into two pieces: one that implements the object and tests it with unit tests; and another one that integrates the object into the existing code
15:13:24 <ihrachys> we will try to merge the former pieces but hold on a bit on the latter
15:13:28 <ihrachys> ok, that's the general stuff
15:13:31 <ihrachys> rossella_s: any specifics?
15:14:21 <rossella_s> ihrachys, not really, apart from korzen's patch https://review.openstack.org/#/c/275790/ that should get merged soonish because it's blocking for the others
15:14:32 <korzen> rossella_s thanks for catching the UT failure!
15:14:56 <rossella_s> korzen, :)
15:14:57 <korzen> and thanks electrocucaracha for double checking! :)
15:15:08 <electrocucaracha> :)
15:15:23 <korzen> I have rebased the patch today
15:15:32 <korzen> over the RBAC object implementation too
15:15:45 <ihrachys> right. we need to prioritize merging pieces that block others - testing, framework enhancements, custom sqla types, ...
15:15:55 <electrocucaracha> korzen, I saw that you include more changes that were not included in patch 8
15:15:56 <rossella_s> ihrachys, right
15:16:04 <ihrachys> oh right, there was RBAC patch that landed recently: https://review.openstack.org/#/c/250081/
15:16:31 <ihrachys> thanks to hdaniel, it's very generic and will hopefully help us once we get to adopting objects for networks
15:16:41 <korzen> electrocucaracha, yes, it was due to the rebase to recent master branch
15:16:56 <electrocucaracha> korzen, gotcha
15:17:36 <ihrachys> there is also a patch from ajo that handles rolling upgrades for qos objects
15:17:38 <ihrachys> #link https://review.openstack.org/#/c/268040/
15:17:51 <ihrachys> I presume it's near getting in, but we'll see
15:18:11 <ajo> there I am :)
15:18:17 <ajo> running unit tests before pushing new patch
15:18:37 <ajo> not only qos objects, any ovo that we will push/pull via that api
15:18:53 <ihrachys> the approach for rpc callbacks is different than what we have in nova and maybe than what we will perceive as a general solution for agent-server interactions, but it's needed for now nevertheless
15:19:27 <ihrachys> some pieces of it will be needed any way
15:19:42 <ihrachys> specifically, automatic RPC version calculation
15:20:42 <ajo> sure, that could be reused for knowing what object versions to use when calling to other pieces in the distributed system
15:20:55 <ihrachys> right
15:20:56 <ajo> if that's what you mean
15:21:03 <ihrachys> ok let's move to the next topic
15:21:05 <ihrachys> #topic Other patches on review
15:21:23 <ihrachys> anything that is worth being mentioned here?
15:21:51 <ihrachys> ok, I have one then :)
15:22:04 <saisriki> I am working on MACAddress type. I have put in the changes in the same file as IPAddress
15:22:12 <ihrachys> saisriki: great
15:22:15 <korzen> I've added the hook to modify the fields before DB operations
15:22:20 <saisriki> custom sqlalchemy type
15:22:24 <korzen> https://review.openstack.org/#/c/281850
15:22:41 <ihrachys> saisriki: any ETA for the code to get into gerrit?
15:23:09 <mhickey> saisriki: whats the url for sqlalchemy type?
15:23:34 <saisriki> I was wondering if it's ok to merge it in https://review.openstack.org/#/c/277558/8
15:24:27 <mhickey> saisriki: I have added PS to this. have you checked them?
15:24:38 <saisriki> I am working locally. I have not uploaded it yet. My question is do I need to create new file or should I put the changes in
15:24:39 <saisriki> sqlalchemytypes.py
15:25:05 <korzen> saisriki, I've created the dependent patch on CIDR: https://review.openstack.org/#/c/285349
15:25:08 <ihrachys> saisriki: please base your change on the first patch
15:25:14 <ihrachys> saisriki: but make it a separate patch
15:25:25 <saisriki> ihrachys: ok.
15:25:30 <ihrachys> saisriki: as for the file itself, you should reuse the existing one
15:25:37 <saisriki> mhickey: what is PS?
15:25:43 <ihrachys> saisriki: patch set
15:25:47 <mhickey> patch set
15:25:53 <saisriki> ok
15:26:08 <mhickey> so if you refresh the patch you will see it is a ps 10
15:26:16 <ihrachys> saisriki: so do you have an estimation when we will be able to start reviewing it?
15:26:43 <mhickey> ihrachys: do you want me to continue with https://review.openstack.org/#/c/277558/?
15:26:47 <saisriki> I will be done in two days, I will push the changes by today EOD
15:27:14 <ihrachys> mhickey: yeah, you were the last one to touch it, right? ;)
15:27:25 <mhickey> ihrachys: yes
15:27:27 <saisriki> mhickey: got it, checked that it is ps 10, Thanks.
15:27:30 <ihrachys> saisriki: ok, let's upload what you have and proceed from there
15:27:39 <mhickey> saisriki: np
15:27:50 <saisriki> ihrachys: ok
15:27:53 <ihrachys> great.
15:28:09 <ihrachys> I wanted to note one bug in ovs agent and how it manages flows.
15:28:11 <ihrachys> #link https://bugs.launchpad.net/neutron/+bug/1514056
15:28:11 <openstack> Launchpad bug 1514056 in neutron "Restarting OVS agent drops VMs traffic when using VLAN provider bridges" [High,In progress] - Assigned to Hynek Mlnarik (hmlnarik-s)
15:28:30 <ihrachys> they say that in some cases we still experience data plane disruption when restarting ovs agent
15:28:38 <ihrachys> so may be interesting to some of team members :)
15:28:53 <ihrachys> there are already some patches on review
15:28:55 <ihrachys> #link https://review.openstack.org/#/c/284639/
15:29:18 <rossella_s> ihrachys, thanks for the pointer
15:29:35 <mhickey> ihrachys: ditto
15:29:54 <ihrachys> ok I guess no more patches to be aware
15:30:03 <ihrachys> #topic Open discussion
15:30:16 <ihrachys> if you had something to raise here, that's the best time
15:31:02 <mhickey> fyi: if you need me just ping me in irc. i had laptop problemsd and was away last week. just getting organised again.
15:31:10 <korzen> does anyone know, if tempest smoke test is doing any integration with DVR?
15:31:41 <ihrachys> not that I know. we probably may want to do some analysis of tests invoked as part of smoke suite.
15:31:42 <korzen> integration, i mean if smoke tests test DVR
15:32:03 <ihrachys> I suspect nothing dvr specific, just general routers. (which will scratch some testing surface for DVR nevertheless)
15:32:50 <electrocucaracha> the only road-block for auto generate the ERD schema is to get an approval of the new dependency in global requests
15:32:51 <electrocucaracha> https://review.openstack.org/#/c/281880/
15:33:30 <korzen> I guess it would be good to test the dataplane disruption in Grenade test...
15:33:33 <ihrachys> electrocucaracha: I see. let me check the list of cores there. We may be constrained by some requirements freeze that may have already occurred for Mitaka.
15:34:23 <ihrachys> korzen: yes. though there are lots of configuration scenarios to test there. vlans may be different from tunnels, provider networks from tenant ones, etc.
15:34:45 <ihrachys> it does not mean we should not take a common one and start tackling in from there :)
15:35:12 <korzen> and before that, we can do the manual check
15:35:27 <korzen> but I would wait on RC
15:35:42 <korzen> I guess it would be possible to validate upgrade on Rc-1?
15:35:46 <korzen> or Mitaka-3?
15:35:55 <ihrachys> I suggest we do it after M3
15:36:12 <ihrachys> since RC in theory should be the final :)
15:36:23 <ihrachys> not that it will, but it's better to have some more time
15:36:53 <ihrachys> maybe a week after M3, when some exceptional pieces may nad
15:36:54 <ihrachys> *may land
15:36:58 <korzen> I wonder when is the cut off date, when we can be sure that no abusive changed got merged
15:38:08 <ihrachys> latest release countdown:
15:38:10 <ihrachys> #link http://lists.openstack.org/pipermail/openstack-dev/2016-February/087590.html
15:38:16 <ihrachys> "Mitaka 3 milestone: Feb 29 - Mar 4"
15:38:31 <ihrachys> so this week
15:39:18 <rossella_s> yay
15:39:33 <ihrachys> :)
15:39:39 <korzen> rossella_s, I have a question about synthetic fields handling - are you going to work on it this week>?
15:39:57 <korzen> this is also kind of important patch
15:39:58 <rossella_s> korzen, that patch is blocked by yours actually
15:40:06 <ihrachys> rossella_s: can we rebase?
15:40:18 <ihrachys> rossella_s: I think korzen's is going to merge once CI votes
15:40:35 <rossella_s> ihrachys, yep , I will rebase mine
15:41:06 <mhickey> rossella_s: needs +w
15:41:27 <rossella_s> mhickey, it needs to be rebased anyway
15:41:39 <mhickey> rossella_s: ok, np, thanks
15:41:44 <ihrachys> btw folks note that zuul is not very happy these days
15:41:58 <ihrachys> I see patches from 10 hours ago that are still in check queue
15:42:00 <rossella_s> indeed
15:42:02 <ihrachys> which is bothersome
15:42:08 <ihrachys> but I bet infra is on it
15:42:10 <mhickey> ihrachys: so avoid recheck?
15:42:23 <ihrachys> mhickey: if you are not sure it will help
15:42:30 <ihrachys> mhickey: ideally, you should read logs first
15:42:38 <mhickey> ihrachys: is it ok to commit ps?
15:42:38 <ihrachys> then decide whether recheck has a chance to help
15:42:46 <ihrachys> mhickey: commit?
15:42:56 <rossella_s> ihrachys, mhickey since the gate is busy I'd avoid rechecks for patches that are not going to be merged soon
15:43:00 <mhickey> ihrachys: git review
15:43:36 <rossella_s> mhickey, yep, we want your patches :)
15:43:44 <ihrachys> rossella_s: that's a good point. let's focus on the bits we plan to merge now (tests, objects framework patches, new objects) and leave other stuff off until gate recovers.
15:44:00 <mhickey> rossella_s: now that I have recovered from disk problems, it maybe possible! :)
15:44:04 <ihrachys> mhickey: yes, we want people not to sit on their patches ;)
15:44:28 <ihrachys> mhickey: especially since discs may fail any time! and you loose your work ;)
15:45:08 <mhickey> ihrachys: lol. losing some of my Irish luck of late! :)
15:45:51 <ihrachys> ok I bet we now know how to behave as good citizens in the community and save resources for the project ;)
15:46:02 <ihrachys> I assume that's all we have for today
15:46:09 <rossella_s> :)
15:46:15 <mhickey> :)
15:46:23 <ihrachys> if not, let's proceed on the #openstack-neutron channel
15:46:27 <ihrachys> thanks bye :)
15:46:27 <ihrachys> #endmeeting