22:14:59 <mlavalle> #startmeeting neutron_drivers
22:15:00 <openstack> Meeting started Thu Feb  8 22:14:59 2018 UTC and is due to finish in 60 minutes.  The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:15:01 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:15:03 <openstack> The meeting name has been set to 'neutron_drivers'
22:15:16 <mlavalle> thanks for pinging me yamamoto
22:15:38 <mlavalle> ihrachys: hi
22:15:43 <mlavalle> armax: hi
22:16:14 <ihrachys> o/
22:16:31 <yamamoto> mlavalle: i was just wondering if i should go back to bed :-)
22:16:43 <mlavalle> sorry, I was looking at the patch from slaweq
22:16:48 <mlavalle> and got distracted
22:19:00 <mlavalle> I don't know if armax will join us for the meeting. worst case I'll update him afterwards
22:19:08 <armax> here
22:19:16 <mlavalle> This is what our dashboard looks like right now:
22:19:34 <mlavalle> welcome, delighted to see you armax
22:19:49 <mlavalle> #link https://launchpad.net/neutron/+milestone/queens-rc1
22:20:32 <mlavalle> Most of it has landed, as you all can see
22:21:23 <ihrachys> I think it's mostly ok / will merge except the MTU bug
22:21:35 <ihrachys> that I don't think will be fixed in time unless we take over
22:21:39 <mlavalle> we had a nasty surprise last Friday: https://bugs.launchpad.net/bgpvpn/+bug/1746996
22:21:40 <openstack> Launchpad bug 1746996 in neutron "bagpipe: IPAllocation DB query failing, engine facade mismatch" [High,Confirmed]
22:22:03 <mlavalle> and ihrachys has done an excellent job wrestling it to submission
22:22:33 <ihrachys> fix on neutron side: https://review.openstack.org/#/c/541512/
22:22:34 <mlavalle> ihrachys: so we are all on the same page, would you let the team know the next steps?
22:22:42 <ihrachys> there is also bagpipe: https://review.openstack.org/#/c/541513/
22:23:47 <mlavalle> this last one just failed unit tests
22:23:55 <ihrachys> also tmorin has https://review.openstack.org/#/c/541775/ for bgpvpn to close potential issues because of mixed engine facade usage but that's not as pressing as the others that are gate-breaking
22:24:36 <ihrachys> I will have a look, but note I just uploaded a new patch set with a fix for bagpipe unit test failures so maybe it's ok too now
22:24:52 <armax> do we want to wait a tad longer for cutting RC1 then?
22:25:21 <ihrachys> while we discuss other things, I rechecked both patches so that we have results after the meeting
22:25:21 <mlavalle> I think that is going to be the case armax
22:25:37 <ihrachys> armax, when did you plan to cut?
22:26:07 <armax> I sent out an email earlier today shouting EOB today
22:26:52 <ihrachys> EOB?
22:27:02 <ihrachys> explanation of benefits?
22:27:04 <manjeets> hi
22:27:14 <ihrachys> nice, I get a bonus!
22:27:59 <mlavalle> armax: can we wait to tomorrow?
22:28:10 <armax> mlavalle: I think so
22:28:33 <ihrachys> yeah tomorrow would be nice. I could clean up / line up all patches till then for OVO bug. otherwise it will slip and it's not good since bagpipe would be effectively broken
22:28:45 <armax> but it depends
22:28:52 <armax> how far off are we from landing those changes?
22:29:09 <ihrachys> I think we are darn close
22:29:19 <mlavalle> I have the same opinion
22:29:57 <ihrachys> if all goes well we may even already have final patches
22:30:00 <armax> can that land today?
22:30:40 <ihrachys> need to wait for ci results. it's pretty late for today probably. but if you folks review it with two +2 and it's just matter of +W after CI passes I can baby sit
22:30:46 <armax> I can have a look at the patches later tonight
22:30:55 <armax> OK
22:31:04 <armax> I can keep an eye on these
22:31:12 <mlavalle> and I don't mind watching it overnight
22:31:25 <mlavalle> ihrachys: if you want, we can take turns
22:31:28 <ihrachys> ok. let's make sure nothing bad sneaks in with the neutron patch. I don't want to break it further :)
22:31:44 <ihrachys> mlavalle, I will code some other thing in the evening so I will keep an eye
22:31:57 <armax> smcginnis’s email states: Before I continue with the countdown info for next week, just a reminder that today, 8th February, is the RC1 milestone
22:32:05 <armax> the patch has been proposed
22:32:24 <ihrachys> armax, maybe if we ask politely..
22:32:32 <ihrachys> anyway, we'll do our best
22:32:34 <armax> but I believe we have time until tomorrow? :P
22:33:04 <armax> I figure the sooner we cut the better
22:33:10 <armax> to avoid any last minute gate crush
22:33:17 <mlavalle> agree
22:33:34 <ihrachys> ok
22:33:42 <ihrachys> as for mtu bug...
22:33:54 <mlavalle> I was about to ask
22:34:18 <ihrachys> I know armax looked at https://review.openstack.org/#/c/499908/ lately
22:34:24 <ihrachys> armax, is it far from completion?
22:34:43 <armax> mine was a nit
22:34:53 <armax> strictly speaking
22:35:07 <armax> but I piggybacked on Slawek’s -1
22:35:19 <ihrachys> ok. btw why is it worth the rc1 target?
22:35:42 <mlavalle> you proposed it
22:35:54 <ihrachys> afaiu to get the bug, you need to configure networks; then change mtu options in underlying infra; restart neutron-servers; delete the network
22:36:06 <ihrachys> yeah but now that we are in pressing mode, maybe we should revisit it
22:36:13 <mlavalle> agree
22:36:22 <ihrachys> the scenario is unfortunate of course, it's a regression from Pike
22:36:36 <ihrachys> arguably if we never found time for it in after Pike, it's probably not as pressing
22:36:49 <mlavalle> I am fine removing it
22:37:17 <ihrachys> ok then let's do it unless others feel strongly. it's better to reduce number of patches we push through right now
22:37:21 <ihrachys> and it's just not ready
22:37:43 <mlavalle> armax, yamamoto: what do you think?
22:38:09 <armax> strictly speaking it’s a blocker
22:38:17 <armax> as there’s no way around it
22:38:19 <armax> is there?
22:38:38 <ihrachys> not change mtu of infra after deployment? :)
22:38:50 <ihrachys> I mean, how often does it happen in real life?
22:39:15 <armax> probably not often
22:39:35 <ihrachys> if we feel that's an important scenario, I can probably take it over, of course that would be rc2 material
22:40:08 <yamamoto> i don't feel it's too important
22:40:22 <armax> it’s a corner case for sure
22:40:25 <armax> but there’s no workaround
22:40:46 <armax> should be safe to punt
22:41:15 <mlavalle> so let's make a point of tackling it early in Rocky and backport it
22:41:35 <ihrachys> + that seems like the right conservative approach
22:41:49 <ihrachys> right as in correct not as in far-right lol
22:42:09 <mlavalle> ihrachys: if you are too swamped, I can take a stab at it
22:42:17 <ihrachys> mlavalle, I would love if you do
22:42:27 <mlavalle> ok, I'll take it over
22:42:41 <ihrachys> the patch is reedip's maybe he will find time to respin it though it took 6 months already so not sure
22:42:55 <ihrachys> I mean took to refine the patch
22:43:16 <mlavalle> I will leave a message stating that if we don't hear from him by next Monday, I'll take it over
22:44:16 <armax> OK
22:44:19 <armax> sounds like a plan
22:45:34 <mlavalle> the other one is https://bugs.launchpad.net/neutron/+bug/1747709
22:45:35 <openstack> Launchpad bug 1747709 in neutron "neutron-tempest-ovsfw fails 100% times" [High,In progress] - Assigned to Slawek Kaplonski (slaweq)
22:46:01 <mlavalle> we have a proposed fix for this one:
22:46:05 <ihrachys> slaweq has a patch but I was hopeful jlibosva gets back from pto and peeks at it before we merge
22:46:15 <ihrachys> so maybe tentatively rc2?
22:46:31 <mlavalle> https://review.openstack.org/542257
22:47:00 <mlavalle> when is he scheduled to come back?
22:47:40 <ihrachys> I think he may be back on Fri but not sure. he took two ptos yesterday and today but not friday (at least not for what I see in internal ml for our team)
22:50:32 <mlavalle> ihrachys: could we merge it now and if it proves problematic or jlibosva wants to add something then we request the rc2?
22:51:34 <ihrachys> mlavalle, I am not sure if it makes sense honestly. there was another patch for the same code that triggered the regression, and it seems that the new patch partially reverts direction of the original
22:51:49 <ihrachys> I am not an expert so I hoped we can get one first before we break even more
22:52:11 <mlavalle> smcginnis: ping
22:52:31 <smcginnis> mlavalle: o/
22:52:40 <ihrachys> we always have a rc2 right? not sure what's the rush with this one if it doesn't have votes, and slaweq is afaiu not even sure it fixes the bug completely. it solved some failures for him but not others.
22:52:54 <mlavalle> smcginnis: we are concluding we are goint to need a rc2
22:53:06 <mlavalle> is that ok?
22:53:13 <mlavalle> what do we need to do?
22:53:39 <smcginnis> mlavalle: That's fine. We try to discourage too many additional RCs, but it is not uncommon to at least have an RC2 due to trailing translations and last minute requirement freeze exceptions.
22:54:12 <mlavalle> smcginnis: thanks. how do I make it official?
22:54:46 <smcginnis> No need to do anything special. You can just cut the RC1 now or as soon as possible if there's some major blockingn reason.
22:55:04 <mlavalle> smcginnis: thanks. will do
22:55:09 <smcginnis> Then if you land anything more important, you can just request another RC2 release as long as it's before the final release date.
22:55:13 <armax> right, rule of thumb I’ve always operated under is
22:55:20 <armax> if you know there’s an RC2 you probably wait
22:55:31 <armax> to cut RC1
22:55:41 <armax> but we’re about to run out of time :)
22:56:13 <mlavalle> so that brings me to the next question
22:56:34 * armax listens
22:56:43 <mlavalle> if we are going to have a RC2 anyway, why don't we cut now
22:56:55 <ihrachys> bagpipe broken?
22:57:01 <armax> right
22:57:02 <mlavalle> oh yeah
22:57:09 <ihrachys> like... if someone would like to test rc1 for them, it's a piece of crap
22:57:11 <mlavalle> I keep making that mistake
22:57:13 <armax> if we can fix things in the next few hours we should wait
22:57:26 <armax> I’d say let’s give tomorrow morning ET as deadline?
22:57:29 <ihrachys> +\
22:57:35 <armax> that should give time to tmorin to catch up
22:57:36 <mlavalle> ok
22:57:43 <mlavalle> we have plan
22:57:51 <yamamoto> ET?
22:57:56 <armax> Eastern Time
22:58:00 <yamamoto> ok
22:58:06 <ihrachys> and we are almost out of time
22:58:29 <mlavalle> ihrachys: I'll ping you around 8pm your time to see how are you doing and see if if you need any help
22:58:38 <ihrachys> ok
22:59:13 <mlavalle> not trying to boss you around. just to see if I you need help :-)
22:59:28 <mlavalle> ok, we are out of time
22:59:34 <mlavalle> thanks for attending
22:59:41 <ihrachys> mlavalle, you are the boss tho :)
22:59:56 <ihrachys> ok ciao, see you all on the other side of rc1
22:59:56 <mlavalle> #endmeeting