15:59:05 #startmeeting kolla 15:59:06 Meeting started Wed Feb 15 15:59:05 2017 UTC and is due to finish in 60 minutes. The chair is inc0. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:59:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:59:10 The meeting name has been set to 'kolla' 15:59:22 #topic rollcall 15:59:26 woot 15:59:27 w00t:) 15:59:28 o/ woot 15:59:28 o/ 15:59:39 o/ woot 15:59:40 o/ 16:00:19 o/ 16:00:26 o/ 16:00:39 lets give few more minutes 16:00:43 o/ 16:00:48 o/ 16:01:59 o/ 16:02:19 o/ 16:02:28 o/ 16:02:36 \o 16:03:19 #topic announcements 16:03:23 o/ 16:03:28 1) We have mascot! https://www.dropbox.com/sh/94pukquiw425ji2/AAAl2wVmm72KHPLCAzR6Uboha?dl=0 16:03:32 #link https://www.dropbox.com/sh/94pukquiw425ji2/AAAl2wVmm72KHPLCAzR6Uboha?dl=0 16:03:40 shifty koala 16:03:46 nifty :D 16:03:50 :D 16:03:53 will we have stickers!? 16:04:03 gema: I think so 16:04:07 +1 16:04:11 gema: vinyl ones I hope 16:04:17 gema: take few for me ;d 16:04:18 hrw: will grab some for you 16:04:19 :D 16:04:20 +2 16:04:28 save me a couple ;) 16:04:36 ;) 16:04:45 will do egonzalez 16:05:01 I hope they'll make plenty 16:05:26 2) Next week is ptg, final schedule is here: https://etherpad.openstack.org/p/kolla-pike-ptg-schedule 16:05:33 I love koala/kolla in the sticker... the green leaf is our gate :) 16:05:47 * portdirect sorry i'm late :) 16:05:55 portdirect: fired 16:05:55 yes, it's as halucynogenic as eucalyptus 16:06:24 endless source of inspiration and confusion 16:06:29 lol 16:06:37 3) next weeks meeting is canceled, I'll send mail 16:06:43 lol 16:07:04 jascott1_: just like running openstack 16:07:19 maybe little more confusion and little less inspiration 16:07:51 so, it seems it's going to be short meeting:) 16:08:14 the best kind 16:08:18 as nobody added anything to agenda, we are close to ptg 16:08:38 there was agenda somewhere? 16:08:38 Jeffrey4l: did we create stable/ocata branch? 16:08:51 #link https://wiki.openstack.org/wiki/Meetings/Kolla 16:08:56 hrw: normally it's there 16:09:01 o/ 16:09:04 kolla ocata branch is created 16:09:10 and kollaansible is WIP. 16:09:16 yay, cool 16:09:19 inc0: sure. now it lists agenda for previous meeting ;D 16:09:31 so we can unblock changes that we wanted to push for pike 16:09:37 so if you wanna to fix bug for ocata release. please file a bug 16:09:51 and fix it in master branch, then backport into ocata branch. 16:09:53 * hrw can say a word or two on debian and !x86-64 if someone wants and there is a time for it 16:09:55 this is necessary. 16:10:07 yes, also please focus on testing 16:10:11 inc0, yep. 16:10:19 at least for kolla project, right now. 16:10:31 kolla-ansible too 16:10:42 kolla-ansible branch is not created. 16:11:03 there are still some issue. i am debugging it. 16:11:07 I'd like to ask kolla-k8s people to test too, to make sure we can move to kolla-ansible ocata for genconfig and stuff 16:11:20 do not know whether need change the kolla-ansible's code. 16:11:27 yeah, but I was talking about testing 16:11:38 kolla-k8s works well. 16:11:46 as far as i know. 16:11:55 Jeffrey4l: what scenarios are normally tested 16:11:58 or required to release? 16:12:07 gate is green. 16:12:11 gema: deploy, upgrade, reconfigure 16:12:14 inc0: thanks 16:12:24 inc0: are there results anywhere? 16:12:30 there is no other complain i heard except inc0 file a bug about genconfig. 16:12:37 but need more info inc0 16:12:55 inc0 we test in the gate every commit kolla-genconfig, although we lack cross-repo ci 16:12:59 well, basically run genconfig on node without running containers or anything 16:13:05 or docker-py for that matter 16:13:11 btw, kolla-k8s has release 0.5.0 tag. thanks for all the hard works ;) 16:13:18 genconfig needs docker-py now, which shouldn't be the case 16:13:42 sdake: we have session for that 16:13:48 inc0, ok. let me check that. 16:14:16 inc0 i think we dont need cross repo gating at all as the plan is to get rid of genconfig dep entirely - that is what the session is about :) 16:14:18 #topic open discussion 16:14:34 Debian! 16:14:34 we have session for both 16:14:38 o/ 16:14:39 i see 16:14:42 yup, hrw go ahead 16:14:47 sdake, kolla-k8s need create create cross gate test with kolla project. 16:14:49 yup corss repo gating ftw :) 16:14:52 I got debian support working again 16:14:53 Jeffrey4l +100 16:15:08 first did tests with jessie-backports as base-tag and went to nice results 16:15:14 cool, it's like resurecting the dead over and over again;) 16:15:20 inc0, i do not think docker-py is required for running "genconfig". i will test this later. 16:15:30 158 binary, 201 source images on x86-64 16:15:41 hrw: how much stickers I'd need to give you to convince you to create gate job for debian? 16:15:52 thats awesome hrw :) 16:15:53 Jeffrey4l, can you give me some advice on this: https://review.openstack.org/#/c/417331/ 16:16:03 but as we (Linaro = me and gema) want to move to stretch I rebased my patches today 16:16:06 Jeffrey4l: issue is - it tries to run handlers 16:16:21 stretch is at 137 binary, 195 source images (also x86-64) 16:16:34 hrw, are you guys planning on using source or packages mainly? 16:16:37 all images which require external repos were not tested 16:16:40 portdirect: source 16:16:51 inc0, but all handler is skipped. 16:16:54 portdirect: we want to have same images based on debian/stretch and centos7 16:17:02 i saw on the ML that debian packaging itself could do with a hand 16:17:19 portdirect: I lack time and knowledge for it. mostly time 16:17:19 portdirect: unfortunately we don't have the manpower for both 16:17:25 hrw: one thing worth looking at is that dockerhub has a lot of debian-based images 16:17:26 duonghq, no idea. does kolla-ansible use tunnel typo? 16:17:36 gema: thats fair :) 16:17:50 duonghq, 0.0.0.0 works but meaningless. 16:17:52 (and hrw as well :) ) 16:17:54 inc0: I am interested only in official ones. which are debian and aarch64/debian at that time 16:18:12 yeah 16:18:15 inc0: non-official iamges can contain anything so I prefer to ignore them 16:18:20 hrw nice work dude :) 16:18:23 +1 to that 16:18:32 which brings us to other stuff: non-x86 architectures 16:18:33 o/ 16:18:40 but we totally need gate for debian 16:18:52 out target is aarch64 (armv8a, 64bit arm, arm64) 16:19:07 Jeffrey4l, I think we can leave the exactly setting to user, w/o trying guess which IP the user want 16:19:08 and we have someone working on power 16:19:14 hrw gating is easy - its mostly copy and paste 16:19:15 I had builds for debian/jessie-backports and ubuntu done 16:19:34 sdake: first I want to get stuff working. 16:19:35 hrw one of the many kolla cores can get you going 16:19:37 hrw agreed 16:19:42 * krtaylor waves hand for pp64le 16:19:44 inc0: we are going to create the gate for that, sdake promised to give me a crash training during coffee and I will pass along the knowledge to hrw :D 16:19:46 krtaylor: love you 16:19:51 hehheh 16:20:03 I can help too 16:20:08 inc0: great, thanks 16:20:14 so !x86 patch is easy: https://review.openstack.org/430940 16:20:23 once we have gate for debian, it becomes first class citizen again 16:20:24 gema wfm depending on when the coffee break is - although we do have a big community of devs that can help as well 16:20:26 we really need to know if things break at any time, so gating is important 16:20:30 it changes a way how repos as handled 16:20:34 sdake: +1 16:20:47 I would also like to bring up health checking of containers and a consistent way of being able to tell if a container has succesfully started and/or is running. 16:20:50 once I get centos7/aarch64 image I will add part which takes care of centos repos as well 16:21:13 klindgren: I have a background in testing, I'd like to be part of that conversation too 16:21:29 there are some issues with !x86 and with debian<>ubuntu stuff but I took care of most of them 16:21:55 klindgren: that would make a great feature in pike 16:21:57 krtaylor: original ppc patch was a bit ugly but after my changes most of it can be recreated 16:22:08 yes, I added a comment to that 16:22:19 I know 16:22:19 it needs to be rebased 16:22:25 not rebased 16:22:27 recreated 16:22:41 I think containers that provide api services would be simple to test. Use the docker HEALTHCHECK command to call into the API 16:22:42 well, yes, and split 16:22:45 will take care 16:23:05 I have my initial learning curve, but I'm on it 16:23:07 but the ones that dont have an API (rpc services) not really sure how to health check those. 16:23:09 klindgren: feel free to submit PoC:) 16:23:20 my laptop has now 1431 kolla generated images 16:23:27 we'd need to work with communities to provide that 16:23:37 have to build most of them on aarch64... 16:23:41 maybe worth submitting few bps for the 16:23:43 m 16:23:49 duonghq, ok 16:24:01 whoever invented --logs-dir deserves a drink 16:24:20 and that's basically all from me 16:24:32 That also reminds me - re: --logs-dir 16:24:32 I'd like to jump in on another topic and say a massive thank-you to bmace (not sure if he's here) on the kolla-k8s front 16:24:51 (and the others that have helped with this): though I've not yet tested his work - I've head that hes got kolla-k8s running out of the gate well, so now we can develop locally! 16:24:52 is their a way to not have the python output from a docker container logged to /var/log/messages 16:24:57 awesome dude - and thanks again! 16:25:03 portdirect: +1 16:25:09 Tired of having double logging. 16:25:44 klindgren: Joshua Harlow wrote it in commit 929d84b143e723735817c88efc5df8d6239c6991 16:25:59 ops, wrong window 16:26:40 klindgren: you can configure docker logging 16:26:47 well - shit. I guess I need to use our newer images then 16:27:08 not really 16:27:12 it's docker conf 16:27:26 https://docs.docker.com/engine/admin/logging/overview/ 16:27:41 anyway, I think we can just move to #openstack-kolla 16:27:44 is it possible to set the logging driver per container? 16:27:47 kk 16:27:52 let's not occupy meeting channel:) 16:27:56 re: ptg schedule - multi-arch is on the waitlist, maybe we can pick a time to do a breakout? or lunch table, beer session... 16:28:08 * hrw waits for 11th March 16:28:11 yeah, or book a room later in the week 16:28:12 krtaylor: gema will be there 16:28:17 portdirect ya bmace rocked it, although it was a team effort of everyone involved in the workflow enviornment work! 16:28:25 gema: that'd work too 16:28:32 or lunch table in Wed;) 16:28:33 inc0: I will try to get us one 16:28:39 inc0: that works too 16:28:41 I am on holidays during ptg 16:28:41 works for me, I'll be there all week 16:28:51 darn 16:28:52 I'll be there during Wed, some others too 16:29:29 +1 lunch on wed then 16:29:38 ok, thanks all for comming, I'll give you 30min back of your life:) 16:29:42 #endmeeting kolla