09:00:23 <flwang1> #startmeeting magnum 09:00:24 <openstack> Meeting started Wed Jun 10 09:00:23 2020 UTC and is due to finish in 60 minutes. The chair is flwang1. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:27 <openstack> The meeting name has been set to 'magnum' 09:00:31 <flwang1> strigazi: are you around? 09:01:10 <jakeyip> o/ 09:01:16 <flwang1> hi Jake 09:01:28 <flwang1> #topic roll call 09:01:30 <flwang1> o/ 09:01:32 <flwang1> brtknr: 09:01:34 <jakeyip> o/ 09:01:49 <brtknr> o. 09:01:57 <brtknr> o/ 09:02:01 <flwang1> jakeyip: thanks for joining us 09:02:10 <flwang1> brtknr: seems we don't have strigazi today 09:02:20 <flwang1> let's have a chat about the hca issue anyway 09:02:44 <flwang1> my main question is, what's the state version now so far? 09:02:49 <brtknr> change topic? 09:02:55 <flwang1> #topic hca 09:03:08 <brtknr> what state version 09:03:46 <flwang1> stable 09:03:47 <flwang1> sorry 09:04:15 <flwang1> i think strigazi also reproduced the thread join issue, right? 09:04:16 <brtknr> umm 09:04:32 <brtknr> flwang1: he used train-stable-2 == ussuri-dev 09:04:47 <brtknr> he hasn't tried train-stable-3 with the fix yet 09:05:22 <brtknr> you havent seen the issue again either... 09:05:25 <flwang1> ok, and the train-stable-3 is the latest victoria-dev? 09:05:32 <brtknr> yes 09:05:37 <flwang1> right 09:05:50 <brtknr> train-stable-3 == ussuri-stable-1 == victoria-dev 09:06:03 <brtknr> i tested train-stable-3 on stein and it works there too 09:06:16 <flwang1> brtknr: we need a place to track this if we decide to do hca release like this 09:06:39 <brtknr> that makes sense 09:06:44 <flwang1> i can't see why it won't work, i just want to understand what the current issue 09:06:52 <flwang1> pls correct me 09:07:14 <flwang1> is the issue strigazi saw the same thread join issue like i reported before? 09:07:42 <brtknr> flwang1: yes 09:07:54 <brtknr> flwang1: have you not been able to reproduce it? 09:08:02 <brtknr> flwang1: have you not been able to reproduce it since? 09:08:20 <flwang1> i didn't check it for every cluster i created 09:08:40 <flwang1> sometimes it doesn't affect the cluster creation 09:08:48 <brtknr> I havent been able to reproduce it in our prod with train-stable-2 or train-stable-3 09:08:55 <brtknr> with VM or baremetal 09:09:01 <flwang1> okay 09:09:04 <flwang1> fair enough 09:09:11 <brtknr> flwang1: do you have nested virt enabled? 09:09:15 <flwang1> no 09:09:17 <brtknr> on your hypervisors? 09:09:25 <brtknr> hmm i wonder if that makes a difference 09:09:33 <brtknr> we have it enabled 09:10:02 <flwang1> good to know 09:10:27 <brtknr> I am not sure what it is for strigazi 09:10:29 <flwang1> i think i'm ok to update those versions as long as we have it documented on our doc 09:11:03 <brtknr> flwang1: what do you want the document to look like 09:11:15 <flwang1> the version mapping 09:11:24 <brtknr> the commit that the version is based on? 09:11:33 <flwang1> yep, it would be nice 09:11:41 <brtknr> ok 09:11:43 <flwang1> otherwise, we can't track the issue 09:12:03 <brtknr> is it enough to document this on storyboard rather than our docs? 09:13:07 <brtknr> I can add a link under each task 09:13:26 <flwang1> hmm... i'd prefer to keep it in tree. how about a Readme.md in the dockerfile path? 09:14:02 <brtknr> Ok but this doesnt make sense for stable branches 09:14:13 <brtknr> Ok but this doesnt make sense for pushing directly to stable branches 09:14:16 <flwang1> brtknr: it's always in master branch 09:14:32 <flwang1> like how we track the addons versions 09:15:04 <brtknr> So we would have 1 commit for changing default on stable branch, another commit to track on master branch? 09:15:15 <brtknr> That doesnt seem coherent 09:15:37 <flwang1> i hear you 09:16:30 <flwang1> can we put the commit id into the description of the container? 09:16:48 <brtknr> Hmm 09:17:22 <flwang1> not very good neither 09:18:10 <brtknr> Not sure how we can provide that to magnum-container-build task on the CI 09:18:57 <flwang1> but you can see my point, right? 09:19:13 <brtknr> yeah I agree that we need some form of tracking 09:19:17 <flwang1> we need a way to track the commit id where the hca being build=t 09:19:45 <flwang1> it would be much easier if the HCA is a standalone repo 09:21:14 <flwang1> not sure if we can use docker label https://docs.docker.com/engine/reference/builder/#label 09:22:55 <brtknr> I think until we have a solution, we can use storyboard to track stable tags against 09:23:07 <flwang1> i'm ok with that 09:23:23 <brtknr> we only really need to tag stable tags, not the dev ones 09:23:37 <flwang1> yep 09:24:09 <brtknr> I remember a while ago strigazi was working on a change to only publish docker image if it doesnt exist 09:24:32 <brtknr> if we made that work, we could control publish of hca tags via the CI 09:24:51 <brtknr> so once published, its immutable 09:25:44 <flwang1> i agree, no matter what tool we use, we should make the image immutable 09:25:53 <brtknr> https://review.opendev.org/#/c/720727/ 09:25:55 <brtknr> this one 09:25:57 <flwang1> for better debugging and tracking issue 09:26:48 <flwang1> i like the idea 09:27:17 <flwang1> we need strigazi's input about this topic 09:27:28 <brtknr> I have left a comment on gerrit 09:29:22 <flwang1> thanks 09:29:56 <flwang1> shall we move to next one? 09:32:59 <flwang1> brtknr: ? 09:33:28 <brtknr> ok all the other topics are things that require review 09:33:52 <brtknr> I am happy with CA rotate one 09:34:08 <brtknr> others i am mostly happy with, just havent had a chance to test them yet 09:34:27 <flwang1> yep, actually i listed them just wanna ask strigazi's attention :) 09:34:36 <flwang1> the ca rotate takes too long 09:35:05 <flwang1> brtknr: thanks for your review, i really appreciate it 09:35:17 <brtknr> ca rotate action takes too long or the review? 09:35:38 <flwang1> the review 09:36:18 <brtknr> :P 09:36:19 <flwang1> if strigazi can buy me a beer or coffee, i won't blame him 09:37:05 <brtknr> flwang1: :) 09:37:13 <brtknr> have you guys met in real life before? 09:37:17 <flwang1> brtknr: pls review this one https://review.opendev.org/#/c/733027/ as well 09:37:47 <flwang1> brtknr: yep, i visited CERN at the Nov 2018 09:38:00 <flwang1> just before Berlin summit 09:38:09 <brtknr> Ah cool 09:38:11 <flwang1> or after i can't exactly remember the date 09:39:15 <flwang1> at catalyst cloud, we're still busy on k8s GA, and we need all of those patches I proposed 09:39:25 <flwang1> so sorry for any pushing :) 09:40:16 <flwang1> brtknr: do you have any topic you would like to discuss? 09:41:03 <brtknr> not today, thanks 09:41:40 <brtknr> weve been working on virtual master baremetal work k8s cluster deployed via magnum 09:42:39 <brtknr> so not been able to be very active upstream recently 09:42:45 <flwang1> brtknr: will you continue your 0 worker cluster work? 09:43:45 <brtknr> yes i should be able to resume that piece soon 09:45:05 <flwang1> cool 09:45:13 <flwang1> alright, i think that's all for today 09:48:06 <flwang1> brtknr: thank you for joining the meeting 09:48:10 <flwang1> #endmeeting