15:59:57 #startmeeting kolla 16:00:00 Meeting started Wed Mar 29 15:59:57 2017 UTC and is due to finish in 60 minutes. The chair is Jeffrey4l. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:04 The meeting name has been set to 'kolla' 16:00:14 #topic roll-call 16:00:17 o/ 16:00:18 o/ 16:00:26 woot o/ 16:00:39 since inc0 is absent today. i will hold today's meeting. 16:00:40 woot o/ 16:00:54 Jeffrey4l, copied 16:01:18 hrw, around? 16:01:33 o/ w00t w00t 16:02:08 o/ 16:02:14 #topic Announcements 16:02:29 1. mitaka and newton next z stream is released 16:02:45 but the images are not pushed into hub.docker.com 16:03:06 we need some volunteer at this point. 16:03:22 o/ 16:03:36 but so many images is time consumer and require a good network and machine ;) 16:03:48 any other new from community? 16:03:59 i can provide a builder if need be 16:04:24 mnaser, cool. thanks. i can add you into kolla namespace in docker after the meeting. 16:04:28 i can also build the images if need be on it 16:04:47 to make sure we're doing it right 16:04:53 i'll provide the commands i'll run 16:04:56 (ill setup the machine as we speak) 16:05:12 ;: 16:05:14 mnaser, any progress for push images to docker automatically? 16:05:22 o/ 16:05:36 Jeffrey4l haven't gotten around revising it based on the feedback i got, lots of people are saying "not so necessary once we get zuulv3" 16:05:39 so thats in a bit of a limbo 16:05:47 wait till zuulv3 happens (when and how i have to check) 16:06:05 or .. do the work which might get phased out when zuulv3 is out 16:06:07 we wait zuulv3 for one year already :( 16:06:13 ok. next 16:06:19 #topic Automated version detection (inc0) https://review.openstack.org/#/c/442923/2 16:06:35 since inc0 is absent, this topic will be pushed into next meeting. 16:06:38 move on. 16:06:44 * hrw 16:06:50 #topic Non-x86 support patchset (hrw) https://review.openstack.org/430940 16:06:55 hrw, yours 16:06:56 zuulv3 is such a bad excuse as zuulv3 is months out mnaser :) 16:06:58 yes 16:07:08 sdake fair enough, thats how i feel 16:07:10 so, you may be aware of my non-x86 work on kolla 16:07:15 why not zuul try micro version ? 16:07:25 Jeffrey4l to stay focused 16:07:34 I rewrote my patchset today to make it easier to merge - no longer one big patchset 16:07:35 sorry. hrw please. 16:08:09 https://review.openstack.org/#/c/451268/ adds base_arch variable to kolla in a away that nothing changes. so far +2 from pbourke 16:08:30 with this in master most of my patches are single changes and can be merged in any order 16:08:35 hrw 1 question, is debian work done? 16:08:50 Jeffrey4l: yes. it is subset of my patches 16:09:09 12:07 hrw@gossamer:diff$ ../stats.sh 16:09:10 centos-binary: 169 16:09:10 centos-source: 206 16:09:10 debian-binary: 142 16:09:10 debian-source: 195 16:09:12 ubuntu-binary: 161 16:09:14 ubuntu-source: 207 16:09:17 thats from x86-64 16:09:19 great. i will enable the debian gate job. 16:09:50 Jeffrey4l: without patchset which update it to use stretch it will mostly fail 16:09:52 question 2: what about for arm build? 16:10:05 or other cpu arch? 16:10:09 I lost connection to my aarch64 build machine but numbers are similar 16:10:12 moment 16:10:21 ppc64le: 16:10:23 debian-binary: 134 16:10:23 debian-source: 187 16:10:23 ubuntu-binary: 143 16:10:23 ubuntu-source: 193 16:10:27 no centos builds here 16:10:42 we can add a experimental jobs. 16:10:58 fails are usually lack of packages or something to fix 16:11:38 question 3: will u provide/prepare third party CI jobs for other cpu_arch? arm/ppc64le 16:12:10 Jeffrey4l: there are plans to make 3rd party CI for openstack jobs on aarch64. no info about ppc64le 16:12:31 got. thanks. this is a great progress. 16:12:37 kurtaylor may know more about ppc64le - I took care of it just as pure curiosity 16:12:55 and ppc64le machine is faster than aarch64 for build tests ;d 16:13:22 roger. 16:13:35 any others? 16:13:50 hrw, or other guys? 16:14:13 seems no. let move on. 16:14:21 #topic post-ptg bps 16:14:22 Jeffrey4l: once base_arch get merged I rebase non-x86 on top and we should have easier jobs 16:14:27 * hrw done 16:14:32 cool. 16:14:48 #link https://etherpad.openstack.org/p/kolla-pike-ptg-blueprints 16:15:12 we have review this many times. but there are still items need pbs. 16:15:30 please open it and leave comments and try to convert it into bps. 16:15:31 thanks. 16:15:40 timebox 20min 16:16:25 i've been trying to find sometime to try out how things would look like running tempest smoke tests as part of gate 16:16:46 and getting rid of the sanity test that is already there (replacing that single one by a tempest one at least) 16:17:48 also, i think it would be interesting (as part of build.py refactor) to make it possible to build images manually as part of an API. 16:18:21 the reason behind it is for testing, instead of tox firing off a build for $world, it could be a list of builds that can fail on their own rather than $world 16:18:49 Jeffrey4l: done my part 16:19:23 mancdaz, run tempest is possible. i tried it. 16:19:42 mnaser, check https://review.openstack.org/402122 ( pinged wrong person again ) 16:20:12 oh thats awesome 16:20:20 this link is 404 https://blueprints.launchpad.net/kolla/+spec/kolla-gui 16:20:26 ill try to pick it up 16:20:55 iirc, there are still some jobs failed. but it is almost there. 16:21:02 some tests* 16:22:29 got some aarch64 stats 16:22:33 Jeffrey4l: corrected https://blueprints.launchpad.net/kolla-ansible/+spec/kolla-gui adding work items for this in progress 16:22:39 centos-binary: 60 16:22:39 centos-source: 190 16:22:39 debian-binary: 129 16:22:39 debian-source: 182 16:22:39 ubuntu-binary: 141 16:22:41 ubuntu-source: 28 16:22:43 got it. thanks. 16:23:28 added numbers to BP 16:23:50 hrw, so for aarch64, you will/want to support all types? centos/ubuntu/debin*source/binary? or only debian? 16:24:22 Jeffrey4l: debian+centos source. 16:25:03 aha, you will use source rather than binary. 16:25:08 we (Linaro) do not support other distros and there are too many binary packages missing 16:25:27 i though you wanna use debian+binary 16:25:30 we may build binary images but more as tests 16:26:59 we have so many jobs about gating. and they also exist long time ago ;( 16:27:23 really need some eyes on gate jobs. ;) 16:27:55 Jeffrey4l: can gate jobs be run on own build machines somehow? 16:28:21 would prefer to first check how they work/run before trying to change anything in official infra 16:28:27 hrw, technically not. it depends on some magic environments and files come from zuul 16:28:32 ;( 16:29:05 but can not did some refactor to make it work i think. 16:29:24 but we can did some refactor to make it work* 16:29:41 ok. long enough, seems there is nothing to update. 16:29:44 let's move on. 16:30:00 #topic Open Discussion 16:30:26 anything want to talk? 16:30:26 i was hoping i could get some eyes on this review: https://review.openstack.org/#/c/435687/ - i tried to ping pbourke for the -1 a few times with no response.. barbican_api doesnt run in stable/newton at all 16:30:42 I plan to make a html table with infos which images got built for which arch (+logs) 16:31:22 as sometimes I loose track which fails where and why 16:31:26 hrw, etherpad may be better? 16:31:30 mnaser: sorry, didnt notice 16:31:32 mnaser: will look now 16:31:36 thank you! 16:31:44 Jeffrey4l: no. python generated html is what I prefer 16:31:49 got. 16:31:55 Jeffrey4l: 3 dirs of text files as input, html as output 16:32:32 mnaser: feel free to pm me if that happens in the future, hard to keep up with all chats :) 16:32:33 Jeffrey4l: sure, etherpad for listing what broke why is good idea 16:32:42 no worries, thanks pbourke 16:33:09 etherpad could tell others the progress ;) 16:33:20 ok. seems nothing to talked. 16:33:28 let end the meeting. 16:33:34 #endmeeting