15:00:09 <bnemec> #startmeeting oslo 15:00:09 <openstack> Meeting started Mon Mar 4 15:00:09 2019 UTC and is due to finish in 60 minutes. The chair is bnemec. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:12 <openstack> The meeting name has been set to 'oslo' 15:00:27 <moguimar> o/ 15:00:29 <bnemec> courtesy ping for amotoki, amrith, ansmith, bnemec, dansmith, dhellmann, dims 15:00:29 <bnemec> courtesy ping for dougwig, e0ne, electrocucaracha, flaper87, garyk, gcb, haypo 15:00:29 <bnemec> courtesy ping for hberaud, jd__, johnsom, jungleboyj, kgiusti, kragniz, lhx_ 15:00:29 <bnemec> courtesy ping for moguimar, njohnston, raildo, redrobot, sileht, sreshetnyak, stephenfin 15:00:30 <bnemec> courtesy ping for stevemar, therve, thinrichs, toabctl, zhiyan, zxy, zzzeek 15:00:37 <stephenfin> o/ 15:00:38 <bnemec> #link https://wiki.openstack.org/wiki/Meetings/Oslo#Agenda_for_Next_Meeting 15:00:39 <hberaud> o/ 15:00:57 <ansmith> o/ 15:01:01 <njohnston> o/ 15:02:52 <bnemec> #topic Red flags for/from liaisons 15:03:41 <redrobot> \o 15:03:47 <bnemec> I think the big thing on our side was that we merged a fix for oslo.cache + py3 + eventlet. 15:03:49 <njohnston> nothing new from Neutron that I am aware of 15:04:30 <bnemec> I was able to get that released before non-client library freeze too (sort of), so that type of environment should work in Stein now. 15:05:22 <bnemec> Otherwise things should be pretty quiet since we're in feature freeze now. 15:05:33 <jungleboyj> o/ Kind of here. 15:06:22 <bnemec> jungleboyj: It's Monday morning. "Kind of here" is about the best anyone can expect. ;-) 15:06:30 <bnemec> #topic Releases 15:06:34 <jungleboyj> :-) 15:06:40 <bnemec> There were a lot of them last week. 15:07:00 <bnemec> I released basically everything that had any pending changes, even if they don't normally require releasing (docs, test changes, etc.) 15:07:30 <bnemec> The reason being that a bunch of our stable branches got broken last cycle because they didn't get some necessary test fixes when we branched. 15:07:41 <bnemec> And that was because they were never released. 15:07:53 <bnemec> Also, we want the stable branch docs to be correct. 15:08:22 <bnemec> Be aware that we are now in non-client library freeze, which includes basically all of Oslo. 15:08:33 <kgiusti> o/ (sorry - late due to snow-pocalypse) 15:08:45 <bnemec> So if there's anything we need to merge, it will require an exception from the requirements team. 15:09:16 <bnemec> Check out smcginnis's last release email for details. 15:09:26 <bnemec> Or just ping me if there's something you think needs to be released. 15:09:52 <bnemec> Otherwise that's probably it for releases this cycle. 15:10:03 <bnemec> Nice work, everyone! 15:10:30 <bnemec> #topic Action items from last meeting 15:10:40 <bnemec> "bnemec to look at hberaud's oslo.cache fix" 15:10:59 <hberaud> https://review.openstack.org/634457 15:11:01 <bnemec> I assume that was the pymemcache one, which I did look at. 15:11:24 <bnemec> So this was done. 15:11:30 <hberaud> yep 15:11:41 <bnemec> I want to discuss it a bit more later, but there's a separate topic on the agenda for that. 15:11:49 <bnemec> "kgiusti and gsantomaggio to investigate OSA rabbitmq_policies" 15:11:59 <bnemec> Any update on this one? 15:12:17 <kgiusti> bnemec: there's a patch outstanding (checking status...) 15:13:15 <kgiusti> bnemec: merged https://review.openstack.org/#/c/639299/ 15:13:18 <bnemec> If a patch was proposed then I'm calling it done. :-) 15:13:23 <bnemec> Even better! 15:13:34 <bnemec> "kgiusti document singleton pattern" 15:13:39 <bnemec> You're up again. :-) 15:13:43 <gsantomaggio> did it! 15:14:29 <kgiusti> bnemec: not done - is this something we can do now that the freeze is in place? 15:15:44 <bnemec> kgiusti: If it's just a doc change we can merge it, but it won't make the Stein docs unless we either have to release oslo.messaging again for a critical fix or we backport it after branching. 15:15:55 <bnemec> kgiusti: Is https://review.openstack.org/#/c/639152/ related to this? 15:16:21 <kgiusti> bnemec: no - that's Something Completely Different 15:16:53 <kgiusti> bnemec: I never understood why rabbit driver opens a new connection for each listener - restored an old comment to the code that explains why 15:17:04 <bnemec> Okay, thanks. 15:17:06 <gsantomaggio> wait 15:17:10 <gsantomaggio> this is not RAbbitMQ 15:17:16 <gsantomaggio> this is Py-AMQP 15:17:33 <gsantomaggio> or oslo-messagging ( I don't know yet ) 15:17:53 <kgiusti> gsantomaggio: yes - I mean the rabbit _driver_ - that's oslo.messaging and it's the code that actually opens the extra conns 15:18:06 <kgiusti> gsantomaggio: something we can fix :) 15:18:19 <gsantomaggio> one of the problem is about the connection 15:18:31 <gsantomaggio> in py-amqp is not thread-safe 15:19:20 <gsantomaggio> so the easier way is to open one connection per thread 15:20:05 <kgiusti> gsantomaggio: yup. that's how the other drivers are designed: one thread dedicated to a URL (socket) 15:20:12 <gsantomaggio> and also another problem is about the basic_ack ( normally should be thread-safe) but in py-amqp is not thread-safe 15:21:22 <gsantomaggio> @kgiusti go it 15:21:26 <gsantomaggio> @kgiusti got it 15:21:42 <kgiusti> gsantomaggio: feast your eyes upon this: https://bugs.launchpad.net/oslo.messaging/+bug/1776492 :) 15:21:43 <openstack> Launchpad bug 1776492 in oslo.messaging "can oslo_messaging support basic.return " [Wishlist,Confirmed] 15:21:53 <bnemec> Okay, so there's more to be done here. 15:22:02 <kgiusti> gsantomaggio: nearly a year old :( 15:22:28 <bnemec> As long as it's on someone's todo list that's good enough for the meeting. :-) 15:22:43 <bnemec> We can certainly follow up on this though. 15:22:50 <bnemec> "moguimar to follow up on status of py35 in the gate" 15:22:55 <bnemec> This was done. 15:23:16 <moguimar> o/ 15:23:20 <bnemec> Let me find a link because it's worth taking a look at. 15:23:56 <bnemec> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-February/003202.html 15:24:16 <bnemec> This is actually going to come up in a later topic too, so we'll move on. 15:24:22 <bnemec> That was it for action items from last week. 15:24:34 <bnemec> #topic Oslo Forum Sessions 15:24:59 <bnemec> I didn't start an etherpad for this because I don't know that we need to have any Oslo-dedicated forum sessions. 15:25:14 <bnemec> I did propose an oslo.limit topic on the Keystone etherpad though. 15:25:41 <bnemec> If you do have something that you think would benefit from discussion at the Forum, please let me know ASAP though. 15:25:52 <moguimar> the Castellan guys are planing to do something I think 15:26:06 <moguimar> redrobot ^ 15:26:25 <bnemec> Cool. 15:26:28 <redrobot> \o 15:27:08 <redrobot> Yeah, There were a few topics in Castellan that need discussion. Not sure if Forum or PTG would be best though? 15:27:46 <bnemec> Yeah, I'll start a new PTG etherpad sometime in the near future. 15:27:59 <bnemec> Maybe I should just go ahead and do that now and we can figure it out from there. 15:28:15 <bnemec> It'll be easier to have the discussion if we have the topics for discussion written down ahead of time 15:28:32 <bnemec> #action bnemec to start PTG/Forum topic etherpad 15:29:28 <bnemec> #topic Bionic Migration 15:29:37 <bnemec> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-February/003129.html 15:30:19 <bnemec> We have about a month before all of the legacy jobs switch over to bionic. 15:30:39 <bnemec> I'm not actually sure if we have much in the way of legacy jobs in Oslo at this point. 15:31:08 <bnemec> Most of our repos are using exclusively templates from the central job repo, and I assume those are all v3-native now. 15:31:39 <bnemec> So it may be that we don't actually have much to do on this. 15:32:06 <bnemec> When I added it to the agenda I was thinking we'd need to push test patches for basically every Oslo repo, but it may be more of an investigation thing instead. 15:32:29 <bnemec> Just verify that our repos aren't using legacy jobs, and only push test patches for the ones that are. 15:32:39 <bnemec> Which may be none. 15:33:14 <bnemec> If anyone wants to dig into zuul configs, be my guest. :-) 15:33:26 <bnemec> Otherwise I'll try to poke at this sometime this week. 15:34:04 <bnemec> Oh, and on a bionic-related note: 15:34:21 <bnemec> I would prefer to hold off on the py35 job removal until after the migration is complete. 15:34:42 <bnemec> The reason being that until that happens, some projects will still be running Oslo code in the gate under py35. 15:35:03 <bnemec> While the unit test jobs don't provide perfect coverage, it's better than nothing and it doesn't cost a lot of resources. 15:36:06 <bnemec> So the main thing is not to merge py35-removal patches until after April 1st. 15:36:25 <bnemec> I'm even going to action that one to highlight it for anyone reading the meeting minutes. 15:36:57 <bnemec> #action Oslo cores to hold py35 removal patches until after Bionic legacy job migration on April 1 15:37:07 <bnemec> No, this is not an April Fool's Day joke. ;-) 15:37:28 <bnemec> #topic pymemcache and oslo.cache 15:37:49 <bnemec> This is mostly for hberaud and kmalloc. 15:38:14 <hberaud> on my side I guess I'm done and I'm waiting kmalloc review 15:38:26 <bnemec> Given that there seems to be some more investigation needed on this in terms of how it works on upgrade, I think we should hold off on this until Train. 15:38:41 <bnemec> Especially since we're in feature freeze and this would add a new dependency. 15:39:10 <hberaud> bnemec: yeah right 15:39:19 <bnemec> Plus we have my awful hack of a "fix" to the existing code that should unblock people using this in Stein. 15:39:48 <hberaud> lol 15:39:58 <bnemec> hberaud: Okay, just wanted to make sure we were on the same page about this. 15:40:21 <hberaud> yeah it's ok 15:40:32 <moguimar> guys, gotta drop sonner, cya o/ 15:40:33 <bnemec> hberaud: I do want to get it in early next cycle so we have lots of time to validate it before the next release. 15:40:44 <bnemec> moguimar: Thanks for joining 15:41:20 <hberaud> sure 15:43:32 <bnemec> #topic Weekly Wayward Review 15:43:38 <bnemec> #link https://review.openstack.org/#/c/610578 15:43:45 <bnemec> I just approved it. :-) 15:44:07 <bnemec> We're pretty limited in what we can merge right now anyway, so we may skip this part of the meeting in future weeks. 15:44:20 <bnemec> A py37 job addition should be okay though. 15:44:30 <bnemec> #topic Open discussion 15:44:35 <bnemec> Anything else this week? 15:44:50 <hberaud> Can some can add this to todo list => https://review.openstack.org/639661 ? 15:46:18 <hberaud> s/some/someone/ 15:47:57 <bnemec> hberaud: How critical is this? I'm not sure we could release it right now even if that merged. 15:48:17 <bnemec> pbr is an independent release project, but we generally still freeze them during feature freeze so we don't accidentally break the world. 15:48:33 <bnemec> Particularly for something like pbr that is used...everywhere. 15:48:48 <hberaud> bnemec: ack 15:50:09 <bnemec> Which is not to say that we can't review the change. It may just have to wait a few weeks until we branch. 15:50:55 <bnemec> I probably need to look closer at the difference between a csv and multi field. 15:51:17 <openstackgerrit> Merged openstack/oslo.cache master: add python 3.7 unit test job https://review.openstack.org/637699 15:51:59 <hberaud> bnemec: ack 15:52:49 <bnemec> Okay, anything else? 15:55:14 <bnemec> Thanks for joining everyone! 15:55:16 <bnemec> #endmeeting