15:00:53 #startmeeting releaseteam 15:00:54 Meeting started Fri Apr 13 15:00:53 2018 UTC and is due to finish in 60 minutes. The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:55 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:55 o/ 15:00:58 The meeting name has been set to 'releaseteam' 15:01:15 #link https://etherpad.openstack.org/p/rocky-relmgt-tracking Agenda 15:01:43 o/ 15:01:44 * fungi is actually around this week 15:01:54 We are currently line ~91 15:02:05 o/ 15:02:19 text in the new etherpad version is bigger by default, we've noticed 15:02:23 Great, good turn out this week. 15:02:33 fungi: I was wondering about that. Thought it was just me. 15:02:44 My failing eyes are OK with that change. :) 15:02:47 totally just you, your eyesight has suddenly improved 15:02:51 hmm, I'm not noticing any difference 15:02:56 Wouldn't that be nice. :) 15:02:58 no diff here 15:03:02 maybe my eyes got suddenly worse? 15:03:05 it may be that it's relative to dpi now or something 15:03:13 Oh, maybe. 15:03:15 ah 15:03:19 appears to be a ~1.2 line vs 1 line 15:03:31 spacing* 15:03:40 yeah, inter-line spacing seems to have increased as well 15:03:50 Looks good on my 34" monitor. 15:03:54 but sorry to derail, this is totally not what we're here to talk about 15:04:12 34? Nice 15:04:18 #topic Review list of exception projects that still tag their own releases 15:04:18 is it not? 15:04:28 ttx: It is very nice indeed. :) 15:04:35 so.... 15:04:54 I wanted us to quickly review the list of "exceptions" and see which ones (if any) we should work on reducing 15:05:19 Like things under openstack-infra should probably not be made compiant 15:05:27 or even compliant 15:05:49 but others are more grey and I could tap into collective wisdom 15:06:08 I copied the list on the etherpad 15:06:18 What is the reason for charms, chef, and helm needing to tag their own? 15:06:27 Well, and rally for that matter. 15:06:39 smcginnis: they probably don't 15:06:49 are may be doing it for lack of release jobs 15:06:56 I could see refstack, and agree on the infra ones. 15:07:01 s/are/and/ 15:07:07 so 15:07:21 Infra is "not openstack" 15:07:38 Ah, so their release needs may be unique enough that we don't have those situations covered right now maybe. 15:08:02 so it does not need to tag through us 15:08:09 Charms 15:08:22 If you look at them they are tagged per Ubuntu releases 15:08:28 I think the thing with infra was they didn't want to wait if they needed a release to fix something in a hurry. Also all of their things are "independent" and we were relaxed about independent projects in the past. 15:08:38 Oh, that makes sense. (charms) 15:09:29 So I guess the question becomes, should we handle tagging for anything in the deployment space 15:10:03 since some of those need tagging to align with the ecosystem they proxy with 15:10:40 We do ansible and puppet. But I suppose these may be distinct enough that we don't need to handle them. 15:11:53 We traditionally did not do Charms or Chef, and haven't been doing Helm yet 15:12:05 OK, let's set that one aside 15:12:28 RefStack -- arguably also not a part of OpenStack but more of a thing that gravitates around it 15:12:40 Rally on the other hand... 15:12:58 is in the openstack-operations bucket and considered a part of OpenSTack (as of today) 15:13:14 Yeah, that one looks the most suspect to me. 15:13:31 Well, the most suspect to me is ec2-api 15:13:48 since it's in the main "openstack" bucket it's as central as it can be 15:13:59 and yet not included in openstack releases I suppose 15:14:06 since we don't tag it 15:14:52 I was just looking at the Teams section. Yeah, ec2-api looks like it should be too. 15:14:53 not even independent or cycle-linked 15:14:59 I guess I'm OK with letting deployment tools tag their own releases. I don't know if any of them would want us to manage the tags for them if we announce that, but are we prepared for a mix (some do, some don't)? 15:15:31 yeah the deployment tools are not my major concern right now 15:15:31 rally and ec2-api feel like things we should be including in the releases site 15:15:37 i foresee infra probably adopting a lot of the same release automation in time, but probably with its own release management team 15:15:39 libraries and openstack bits are 15:15:52 dhellmann: +dragonflow 15:16:03 yeah, I see no reason to make exceptions for those 15:16:08 all the others are more corner cases 15:16:20 like openstack/manila-test-image 15:16:31 Was there some talk of rally moving out of OpenStack proper into a more general Python security thing? 15:16:32 not really part of the "product" 15:16:43 smcginnis: that was bandit, not rally 15:16:49 Oh, right! 15:17:07 smcginnis: rally just moved openstack-specuific stuff into a plugin to be more generally applicable 15:17:10 though rally has recently decomposed its openstack-specific bits and moved them into their own repo so that it can be a more generalized testing framework, according to that recent ml thread 15:17:19 right that ^ 15:17:46 So the other one I'd like to prioritize is openstack/yaql 15:17:53 since afaict it's a lib 15:17:57 yes 15:18:10 that's part of mistral, iirc 15:18:22 it is 15:18:26 and maybe openstack/networking-powervm and openstack/nova-powervm, if we do dragonflow 15:18:27 murano 15:18:32 I thought we had all of the libraries covered already 15:18:46 not mistral 15:18:52 Ah, I didn't think that was used by other deliverables, but if others are, that should be part of the deliverables. 15:19:09 it's in global reqs yes 15:19:29 So I created tasks on that story: 15:19:37 I think we want anything installed in a production deployment to be listed on releases.o.o 15:19:51 * jungleboyj sneaks in late 15:20:06 https://storyboard.openstack.org/#!/story/2001831 15:20:39 Hey jungleboyj 15:20:58 ttx: Great, those tasks look good to me. 15:21:02 And now we have somewhere to track it. 15:21:11 I just adjusted them 15:21:17 Trying to prioritize them 15:21:39 @! 15:21:40 <_pewp_> jungleboyj \( ・_・) 15:21:56 I kept only 5 for the time being 15:22:03 (you might need to refresh page) 15:22:15 So we should contact those teams and make sure they are aware of it (and make sure they don't have a good reason for not changing things). 15:22:22 We traditionally did not do Charms or Chef, and haven't been doing Helm yetec2-api dragonflow rally powerVMstackers and yaql 15:22:31 err 15:22:37 ec2-api dragonflow rally powerVMstackers and yaql 15:22:50 That list is a good start. 15:23:19 OK, if you contact any of those teams, don't forget to put your name on that task 15:23:25 so that we don't duplicate work 15:23:37 I can post something on the ML tagging these teams. Unless someone else wants to take that task. 15:23:48 go for it 15:23:51 d0ugal seems to be here, we could tackle yaql now 15:23:59 That is all I had 15:24:13 dhellmann: What needs to be tacked? :) 15:24:28 tackled* 15:24:49 d0ugal : we would like to add yaql to the list of repos with releases managed through openstack/releases. That means an ACL change and importing the history. I can do that for you if you don't have an issue with it. 15:25:34 it's not really clear why it isn't already being handled that way 15:25:48 i'm curious why we're asking d0ugal though 15:25:56 as it's a murano deliverable, not mistral 15:26:03 Yeah, that is the confusion 15:26:08 Thanks fungi - I was just trying to type it up 15:26:16 oh, is it? 15:26:20 Mistral uses YAQL only 15:26:20 https://governance.openstack.org/tc/reference/projects/murano.html#yaql 15:26:20 why did I think it was mistral? 15:26:28 ok, nm then 15:26:30 sorry, d0ugal 15:26:34 dhellmann: np 15:26:41 * fungi tried to correct that misconception up at 15:18 15:26:53 I would be happy to help more with YAQL as we are heavy users... but that is another matter :) 15:27:03 yeah, I thought when d0ugal said "it is" he was saying it was owned by mistral, not just used 15:27:10 my bad. 15:27:22 * d0ugal steps aside 15:27:45 zhurong doesn't seem to be in channel, but this may also be a bad time to get ahold of them 15:27:52 probably 15:27:59 We can cover it on the ML. 15:28:00 I'll wait for smcginnis' email 15:28:22 ttx: AOB? 15:28:43 #topic Availability next week 15:29:21 yeah won't be around 15:29:22 Ah - absent du bureau 15:29:37 Will be drinking Tuscan wine with flaper87 in Florence 15:29:44 So ttx and I are out Thursday and Friday. 15:29:46 Nice! 15:29:52 AND not even ashamed 15:30:14 I may be around, but I will be chaperoning a group of high school students in Chicago, so not sure on my availability. 15:30:21 So don't want to commit to anything. 15:30:24 that sounds like a full time job 15:30:33 dhellmann: Do you want to run the meeting next week? Or should we just skip. 15:30:46 we can probably skip the meeting if it's just me and fungi. I know how to reach him if we have issues. 15:30:49 It is Rocky-1, so not sure. 15:31:10 sure, i'll be around 15:31:11 But yeah, we can probably sync up async in our normal channel. 15:31:11 if people have issues, they usually come to the channel rather than the meeting anyway 15:31:28 though we do seem to have a nice crowd here today :-) 15:31:41 dhellmann: drinking with flaper87 is indeed a full time job 15:31:59 heh 15:32:00 i have no opinion on whether it's necessary to hold the meeting, but happy to help out with any issues which arise either way 15:32:11 Let's plan to skip. We can leave it up to you if you decide to hold one based on what the need ends up being. 15:32:17 wfm 15:32:24 * flaper87 is proud to be a job provider for ppl around the world 15:32:30 #info Likely skipping next weeks meeting unless some compelling need comes up 15:32:35 flaper87: :) 15:32:53 #topic Rocky-1 actions 15:33:08 Just thought we should do a quick update ahead of next weeks milestone 15:33:22 I actually think we are in great shape. At least better than we were for queens at this point. 15:33:37 So (knock on wood) I don't expect next week to be too crazy. 15:33:53 making headway while we can, because the switch from xenial to bionic will probably set people back 15:33:57 I have a foggy mind today, which I blame on a cat that decided 6:00 was a good time to turn the bed into a race track. What did we decide about applying tags for teams? 15:34:15 for teams that do not submit them, that is 15:34:39 I actually meant to bring that up as it was one of the things I moved to SB. 15:34:54 I _think_ we said we force tagging if it was missed. 15:35:15 so maybe on friday around our meeting time? 15:35:19 yes I think we said that would tarin people to expect us to step in 15:35:23 train 15:35:48 downside being for milestones I expect most would just wait for us to step in 15:36:10 can't remember if we said that we'd do it for milestone-1 though 15:36:10 I do think we need to track which ones we initiate so if we end up forcing it this cycle, we kick them out of the common release next time around if they do not step up. 15:36:32 ok, so it would still be condidered a fail 15:36:33 Maybe good to just force it starting with milestone 2? 15:36:43 yeah... 15:36:48 We can miss m-1 15:36:51 ok, that sounds good 15:36:56 use m-2 as a reminder 15:37:03 so that we can step in m-3 15:37:17 and if people missed too often we have "the discussion" with them 15:37:23 ttx: Wait, force only on 3? Or 2 and later? 15:37:31 break a few fingers 15:37:35 2 and later 15:38:10 I need to run in a minute 15:38:15 OK, so the plan is: miss milestone 1 - public shaming on the ML, miss milestone 2 - public shaming an ML and we force a release, etc. 15:38:36 I'm fine with that 15:38:45 wfm 15:38:46 And if they miss all, they are on "the list" and if they miss milestone 1 in the next cycle they are out. 15:38:50 I'm going to write that up in a patch to the readme 15:39:08 dhellmann: Want to take my SB task for that? 15:39:15 sure 15:39:17 yes our aging brains are not working that well as a memory device those days 15:39:32 https://storyboard.openstack.org/#!/story/2001847 15:39:33 ttx: it's all that drinking with flaper87 15:39:38 Haha 15:40:05 * fungi hasn't been drinking with flaper87 lately, so not sure what my excuse is 15:40:36 must be rum 15:40:44 Or sake. 15:40:51 ok got to run 15:40:54 ttyl 15:40:59 OK, anything else we need to discuss for R-1? 15:41:01 both probably 15:41:01 ttx: o/ 15:41:13 don't assign too many work items to me while I look the other way 15:41:22 * flaper87 is happy to be blamed for these things 15:41:25 I READ LOGS 15:41:25 * smcginnis fires up SB 15:41:37 fungi: we can fix that in YVR 15:41:49 flaper87: you're happy to be blamed for everything 15:42:05 flaper87: you're on ;) 15:42:08 ttx: it's tough but someone has to do the job 15:42:21 can someone remind me how to link to a storyboard story in a git message? 15:42:25 "Story-ID"? 15:42:33 Story: #NNNNN 15:42:37 Task: #NNNN 15:42:46 thanks 15:42:48 capitalization, spacing and the # are optional 15:43:00 we have documentation somewhere ;) 15:43:04 fungi: Is it both story and task? 15:43:15 Or just task is enough? 15:43:25 #link https://review.openstack.org/561258 15:43:26 patch 561258 - releases - document that we will force tag milestones 15:43:41 dhellmann: You're quick! 15:43:50 I didn't want to forget to do it ;-) 15:44:00 story makes its-storyboard leave a comment on the story with a hyperlink to the review (or it should, but we broke that and i'm trying to hunt down how we should fix it). task changes the status of a single task 15:44:25 oh, I guess I could add the task too 15:44:35 we separated story commenting from task status updating because by composing the two (or not) you can achieve different desired effects 15:45:11 like you may want to have a review leave a comment on a story even if there's no task associated with it, or you may want to update multiple tasks 15:45:24 makes sense 15:45:26 Makes sense. 15:45:30 :) 15:45:36 OK, anything else for today? 15:46:03 #link https://docs.openstack.org/infra/manual/developers.html#working-on-bugs Working on Bugs 15:46:14 is where we have it documented in the developer workflow 15:46:17 this week at some point tonyb mentioned in a review comment that we need to be able to show the support status for a deliverable in a series 15:46:23 I opened a story for that 15:46:38 #topic Open discussion 15:46:48 dhellmann: I missed that. 15:46:52 but I can't figure out how to make SB open it in a way that I can past the link 15:47:02 here we go 15:47:03 #link https://storyboard.openstack.org/#!/story/2001852 15:47:43 So for extended maintenance, being able to show the state of a stable deliverable. 15:47:46 that web ui is too smart; it doesn't work with right clicking to open things in new windows 15:47:50 right 15:48:04 the point is that some things won't actually have support so we can no longer say that "ocata" is under EM 15:48:09 although that might be the default for ocata 15:48:33 A repo tag? Or another line in the deliverable file of that cycle? 15:49:02 I was thinking a new field in the deliverable file, with a default that comes from something built into the rendering code 15:49:33 I put it on the backlog for now 15:49:49 OK, sounds good. 15:50:07 And then we'll need some update on the web page to reflect that. 15:50:13 right 15:50:44 fungi : after you figure out what's going on with commenting on stories, maybe you can also help with https://storyboard.openstack.org/#!/story/2001806 15:50:53 I think the top level releases.o.o can still show the overall phase that a release is in, but then going into that release, each project needs its own status. 15:51:07 we want comments on stories or tasks or whatever when a change is released, like we have with LP 15:51:26 smcginnis : yeah, we'll have to figure out how to represent that but you're probably right 15:51:55 dhellmann: sure, that shouldn't be too hard 15:52:34 one of these days I'll get back around to try to release the storyboard client 15:52:35 the bit with gerrit not commenting on stories is harder to track down due to the opaque nature of its inferred configuration and being implemented in java (as a gerrit plugin) 15:52:57 oh, yeah, this other thing should be easy in comparison 15:56:54 Any final things? 15:57:06 smcginnis, ttx: it looks like you opened stories. Maybe add them to the backlog column on https://storyboard.openstack.org/#!/board/64 as well? 15:57:39 nothing from me 15:57:47 I needed to look into that next. Too bad it doesn't automatically pick those up. 15:58:10 it can pick them up if you do an automatic lane based on a tag name or something 15:58:11 hmm, I suppose I could set the board up that way 15:58:22 I can get mine added. 15:58:33 yeah, we could just do it by repo and status or something 15:58:43 that makes it harder to separate backlog and todo, though 15:58:56 We can follow up on that later. 1 minute left. 15:58:59 k 15:59:12 right. using story tags can make some of that viable depending on what you want 15:59:13 Guess that's it for the meeting then. Thanks everyone. 15:59:34 thanks smcginnis!@ 15:59:41 thanks smcginnis! 15:59:48 #endmeeting