09:01:29 <flwang1> #startmeeting magnum 09:01:30 <openstack> Meeting started Wed Apr 22 09:01:29 2020 UTC and is due to finish in 60 minutes. The chair is flwang1. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:01:31 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:01:33 <openstack> The meeting name has been set to 'magnum' 09:01:40 <flwang1> #topic roll call 09:01:40 <cosmicsound> hi 09:01:46 <strigazi> o/ 09:02:34 <strigazi> ttsiouts: brtknr: ping 09:02:40 <flwang1> o/ 09:02:40 <brtknr> o/ 09:02:56 <brtknr> cosmicsound: ? 09:02:56 <flwang1> thanks for joining, guys 09:03:26 <flwang1> let's go through today's agenda 09:03:35 <flwang1> https://etherpad.opendev.org/p/magnum-weekly-meeting 09:03:45 <flwang1> #topic CI 09:03:48 <flwang1> brtknr: ^ 09:03:57 <flwang1> anything we need to discuss about CI? 09:04:02 <strigazi> yes 09:04:10 <strigazi> there are four patches 09:04:17 <strigazi> maybe help with some context 09:04:55 <strigazi> are they ready for input from non zuul people? 09:04:56 <brtknr> 5 patches, including the one in devstack: https://review.opendev.org/#/q/topic:magnum-tempest-plugin-tests-api+(status:open+OR+status:merged) 09:05:07 <brtknr> yep its all ready 09:05:14 <brtknr> as of this morning :) 09:05:20 <flwang1> strigazi: will those patches can run our tempest in CI? 09:05:23 <strigazi> more ci 09:05:32 <ttsiouts> o/ 09:05:40 <ttsiouts> sorry for being late 09:05:44 <strigazi> brtknr: is there an order? 09:06:21 <flwang1> based on my understanding, we still need the nested virt to run magnum tempest to deploy a k8s cluster 09:06:43 <brtknr> strigazi: there is order based on depends-on 09:07:10 <strigazi> i'll figure it out 09:07:25 <dioguerra> o/ 09:07:37 <flwang1> brtknr: strigazi: please correct me if i'm wrong 09:07:45 <brtknr> flwang1: this is the function-api test only, we can worry about functional-k8s test later 09:08:01 <flwang1> not really 09:08:22 <brtknr> although functional-k8s works for me locally 09:08:26 <strigazi> flwang1: i don't epxect the CI to work for spinning clusters. I'm happy to take the patches to conform with zuul 09:08:42 <flwang1> strigazi: that's my question actually 09:08:49 <flwang1> i don't think it could work 09:09:25 <flwang1> though it may be able to cover some very simple api test 09:09:35 <strigazi> but in any case, as brtknr mentioned in storyboard, we need this for zuul 09:10:09 <flwang1> strigazi: agree, good to have 09:10:44 <strigazi> I'm a bit lost with them but I'll figure it out 09:11:12 <strigazi> I'm covered for this topic, thx brtknr 09:11:57 <flwang1> #topic Scrape internal kubernetes components - https://review.opendev.org/#/c/721284/ 09:12:06 <flwang1> anything we need to discuss for this one? 09:12:50 <dioguerra> There are this 3 patches id like to merge before i move with the rest 09:13:18 <dioguerra> as they will conflict. 09:14:04 <flwang1> dioguerra: is there any dep between there? 09:14:28 <flwang1> do you have a recommendation about the review order? 09:15:28 <dioguerra> Not dependency but they interact in the same file. 09:15:57 <brtknr> strigazi: i have made a list of dependencies in the meeting notes 09:16:06 <brtknr> for the ci 09:16:16 <flwang1> brtknr: thanks, that's helpful 09:16:31 <dioguerra> order can be anyone i guess. not one specific but leave https://review.opendev.org/#/c/721592/ for last 09:16:57 <dioguerra> and this one first https://review.opendev.org/#/c/720196/ 09:17:06 <ttsiouts> dioguerra: just rebase the changes on top of each other 09:17:36 <ttsiouts> to give clear indication to the reviewers as to what has to be reviewed in what order 09:17:53 <ttsiouts> this way you won't have conflicts between your changes 09:18:01 <dioguerra> ok 09:18:13 <flwang1> ok, except those 3 patches related to prometheus , there are not much on the agenda 09:18:29 <flwang1> brtknr: strigazi: anything else you want you discuss? 09:18:43 <strigazi> git review -x <six digts from url> # with this you can cherry-pick 09:18:56 <strigazi> and stack the patches 09:19:11 <flwang1> we're on RC1, we need to have a general idea about the features/fixes we want for U cycle 09:19:35 <strigazi> flwang1: let's discuss override-lables 09:19:48 <strigazi> *labels 09:21:15 <strigazi> let's discuss U after? 09:21:37 <strigazi> flwang1: ^^ 09:22:14 <flwang1> sure 09:23:24 <strigazi> brtknr: ttsiouts: is the PoC the final proposal? Shall we agree and move one with it? 09:24:21 <brtknr> i am happy with the PoC personally! 09:24:26 <ttsiouts> strigazi: I have no problem if we all agree it could be it 09:24:41 <ttsiouts> strigazi: what do you think? 09:24:56 <brtknr> I am happy to hear reasons it could present problems later 09:24:59 <flwang1> can you please help me understand the extra overriden_labels here? 09:25:24 <brtknr> flwang1: have you played around with the PoC patches yet? 09:25:31 <strigazi> I'm ok with PoC, the sooner we have it merged the better. 09:25:32 <flwang1> no 09:25:55 <brtknr> flwang1: it may make sense when you try it, it didnt make intuitive sense to me either 09:25:56 <ttsiouts> flwang1: the overridden_labels is just a field that is generated on cluster/nodegroup show 09:26:35 <ttsiouts> flwang1: it contains the diff between the cluster and cluster_template labels or nodegroup and cluster labels 09:27:05 <flwang1> ttsiouts: ok, then that's the answer for my question in comments 09:27:18 <ttsiouts> flwang1: yes 09:27:57 <ttsiouts> I proposed to PoC because this way we don't have to change the schema 09:28:14 <flwang1> i'm OK with it 09:28:42 <flwang1> my question for you guys is, are we going to get it in U? 09:29:03 <strigazi> as we want 09:29:11 <strigazi> +1 from me 09:29:15 <brtknr> should we hold back the release until it merges? 09:29:33 <brtknr> ttsiouts: how long do you think it will take you polish it off? 09:29:42 <flwang1> brtknr: we don't have to hold back the rc1 actually 09:30:25 <strigazi> +1 ^^ 09:30:48 <flwang1> i'm ok to get it in U, my only requirement is pls updating the api ref 09:30:54 <ttsiouts> brtknr: I can quickly fix some things by tomorrow and have it ready by Friday 09:31:21 <ttsiouts> brtknr: I mean for proper review 09:31:21 <strigazi> maybe do the spec first? 09:31:37 <ttsiouts> strigazi: oh yes. forgot about the spec 09:31:37 <strigazi> or in parallel 09:32:06 <ttsiouts> I could start from the spec maybe to agree on the details. is that ok? 09:32:17 <strigazi> +2 09:32:21 <flwang1> +1 09:32:24 <brtknr> sure 09:32:27 <ttsiouts> great! 09:32:31 <strigazi> Just to have it written 09:32:47 <strigazi> and not look at code documentation :) 09:32:55 <ttsiouts> thank you guys for your time! 09:33:22 <brtknr> thank you! 09:33:43 <brtknr> i have 1 more topic for discussion, helm v3 09:35:17 <flwang1> brtknr: pls continue 09:35:54 <brtknr> do you think the best thing is to have a new namespace for helmv3? 09:36:00 <strigazi> yes 09:36:16 <brtknr> ok 09:36:27 <brtknr> i will try and update the patch today 09:36:45 <strigazi> this was it? 09:37:48 <brtknr> have you guys had a chance to test it yet by any chance? 09:38:20 <strigazi> no me 09:38:27 <flwang1> your v3 patch? not yet 09:38:44 <strigazi> is it ready for testing? 09:38:52 <brtknr> the other option i was thinking was to rename tiller namespace to HELM_NAMESPACE defaulting to magnum-helm 09:39:15 <strigazi> let's not do renames :) news things are better :( 09:39:34 <flwang1> +1 for new namespace 09:39:44 <brtknr> ok 09:39:52 <brtknr> do we want it to be configurable via a label too? 09:40:24 <strigazi> both work for me 09:40:40 <brtknr> ok 09:40:54 <brtknr> that gives me enough to get going 09:41:13 <brtknr> yes the PS works for helmv3, just not polished yet 09:41:20 <brtknr> so fine for testing 09:41:39 <flwang1> let's move on? 09:42:18 <strigazi> +1 09:43:00 <flwang1> strigazi: https://review.opendev.org/#/c/718296/ docker storage for FC , are you OK without the support of devicemapper for FC? 09:43:09 <strigazi> +1 09:43:39 <flwang1> and I can propose another patch to add a separate release note and warning in logs for the deprecation of devicemapper 09:43:49 <flwang1> then we can change the default value in V cycle 09:44:01 <brtknr> +2 09:44:17 <flwang1> thank you 09:44:28 <brtknr> when should we have a virtual PTG for V cycle? 09:44:42 <strigazi> s/when// ? 09:44:54 <flwang1> when for what? a following patch? 09:44:59 <flwang1> i will do it tomorrow 09:45:13 <brtknr> should we have a virtual PTG for V cycle? 09:45:26 <flwang1> sure 09:45:43 <strigazi> I don't think we need to have a long one 09:45:56 <flwang1> i think 2hours max 09:46:12 <strigazi> Do we expect anyone that is not here now to attend? 09:46:26 <flwang1> hmm, yes 09:46:37 <strigazi> we do? who? 09:46:46 <flwang1> i'd like to invite all exisitng adopters 09:47:00 <brtknr> Perhaps we should announce it on ML 09:47:05 <flwang1> we will 09:47:08 <strigazi> but will they join? :) 09:47:14 <brtknr> we'll find out 09:47:16 <flwang1> let's try 09:47:34 <flwang1> let's try 6 May? 09:47:35 <brtknr> we have a lot of shadow users 09:47:47 <flwang1> after 2 weeks 09:47:55 <flwang1> UTC9:00- 11:00AM? 09:47:57 <brtknr> sounds good 09:48:19 <strigazi> I don't want to be too negative, let's do it 09:49:31 <flwang1> strigazi: the worst case is they won't join, that's ok 09:49:58 <flwang1> strigazi: anything you want to talk about the the RC? 09:50:33 <strigazi> flwang1: actually no, we can move with rc1 and add what we *need* and want to add 09:50:41 <brtknr> i havent tested the master branch recently, hopefully nothing is broken 09:50:45 <flwang1> cool 09:51:29 <flwang1> for https://review.opendev.org/#/c/714021/ the dashboard 2.0.0, if there is no GA, until we have to GA, are you guys happy with a rc version now? e.g. v2.0.0-rc7? 09:51:40 <strigazi> +1 09:51:54 <strigazi> is it on by default? 09:52:02 <flwang1> yes 09:52:06 <flwang1> IIRC 09:52:08 <strigazi> but we can disable with the label 09:52:12 <flwang1> sure 09:52:16 <strigazi> fine for me, for gor it 09:52:21 <strigazi> fine for me, go for it 09:52:44 <flwang1> we can easily update the version as long as it GA later, not a big deal TBH 09:52:47 <flwang1> ok, cool 09:52:53 <flwang1> i will update the patch 09:53:01 <strigazi> Last thing from me, when you are done 09:53:08 <flwang1> i'm good 09:54:37 <strigazi> recently I abandoned all patches openstack/magnum (not updated for 30d) and brtknr abandoned some patches in openstack/python-magnumclient when I mentioned it. 09:55:19 <strigazi> I'll do the same with all tasks in storyboard if there is not udpate in the last 30d if you agree 09:55:38 <flwang1> strigazi: i'm ok with that 09:55:53 <brtknr> i think 30d might be too short, maybe 2m? 09:55:58 <brtknr> 60d 09:56:14 <strigazi> I'll do it right now. 09:56:46 <strigazi> I prefer 30d, I don't think 3 people have the capacity to review all that properly. 09:57:10 <strigazi> We are not vetoing against anything 09:57:11 <flwang1> i generally only review the code patches :) 09:57:21 <brtknr> mark them as invalid? 09:57:54 <strigazi> if there is anything iportant, people can speak up. Maybe flwang1 we'll manage to review storyboard as well if the backlog is small 09:58:28 <brtknr> strigazi: what will we set the status to ? 09:58:31 <brtknr> invalid? 09:58:42 <strigazi> yes. I have a feeling that you will start doing it. 09:58:47 <brtknr> there is no "stale" label lathough it would be useful 09:58:50 <flwang1> strigazi: +1 09:59:07 <brtknr> i wont touch it 09:59:09 <openstackgerrit> Feilong Wang proposed openstack/magnum master: [k8s] Fix no IP address in api_address https://review.opendev.org/721791 09:59:12 <brtknr> i promise 09:59:47 <strigazi> I will share the code of the client 09:59:50 <flwang1> ok, anything else? 10:00:02 <flwang1> we're running out of time 10:00:14 <strigazi> i'm good 10:00:20 <flwang1> brtknr: ? 10:00:28 <brtknr> nope all good 10:00:34 <flwang1> thank you, guys 10:00:42 <flwang1> #endmeeting