15:00:06 #startmeeting releaseteam 15:00:07 Meeting started Fri Sep 21 15:00:06 2018 UTC and is due to finish in 60 minutes. The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:11 The meeting name has been set to 'releaseteam' 15:00:26 o/ 15:00:27 Ping dhellmann, ttx, annabelleB 15:00:31 Ohai 15:00:31 o/ 15:00:34 o/ 15:00:35 o/ 15:00:51 Wow, everyone is punctual today. 15:00:59 #link https://etherpad.openstack.org/p/stein-relmgt-tracking Agenda 15:01:04 Line 31 15:01:18 #topic Backlog discussion from the PTG 15:01:33 #link https://storyboard.openstack.org/#!/project_group/73 Backlog 15:01:44 We started going through ^ at the PTG but ran out of time. 15:01:54 or steam 15:02:04 choo choo 15:02:04 A little of both. 15:02:37 Did we leave off on stable deliverables or external releases? 15:02:41 can we take elements on that storyboard freely? 15:02:42 * fungi was not punctual :( 15:02:53 fungi: Close enough. ;) 15:03:02 smcginnis: we stopped at stable 15:03:12 evrardjp: If there is anything on there you would like to work on, feel free to grab it. 15:03:22 judging by lastupdated date 15:03:39 * smcginnis notes ttx is much more observant today than himself 15:04:08 Track externally-released deliverables is still on my lap 15:04:09 #link https://storyboard.openstack.org/#!/story/2002624 Track externally-released deliverables 15:04:15 I have a plan 15:04:45 OK, anything you need input on? Or shall we move on? 15:04:49 nope alls et 15:05:17 #link https://storyboard.openstack.org/#!/story/2001936 Reno traceback when no release between 2 tags 15:05:24 will comment so that it's updated 15:05:30 ttx: Thanks 15:05:44 dhellmann: Is the reno issue something you have any chance to look at? 15:06:02 I haven't. It looks like I asked a question with no response. 15:06:20 Ah, see that now. 15:06:25 I suspect there's an issue using --earliest-version and --version together 15:06:33 that first option isn't needed if the second is used 15:06:50 maybe the fix is to make those mutually exclusive 15:07:01 So keep around as a low priority thing to checkout out some day? 15:07:17 dhellmann: +1 15:07:32 smcginnis: agreed too 15:07:40 yeah 15:07:42 #link https://storyboard.openstack.org/#!/story/2001935 cache file needs to contain all branches 15:07:50 Another reno one. 15:08:01 that's in process and also a low priority 15:08:13 that has to do with building the cache file for release notes for a package 15:08:27 Patch up for it too, just with some issues yet. 15:08:32 So in the works at least. 15:08:57 #link https://storyboard.openstack.org/#!/story/2001921 enhance version check for xstatic packages 15:09:09 We had some issues in our validation for these. 15:09:21 that's done 15:09:27 I added the details to the story 15:09:41 OK, I was thinking this was something already addressed. 15:09:57 Last one... 15:10:01 you fixed it :-( 15:10:02 I mean :-) 15:10:08 * elbragstad walks in late 15:10:17 #link https://storyboard.openstack.org/#!/story/2001881 Create a new-deliverable command 15:10:23 Seems like another low priority one. 15:10:29 yeah 15:10:41 we've been doing OK with creating the stub files by copying existing ones for now 15:10:46 Since these are fairly infrequent and easy enough to do through the time honored copy/paste method. 15:10:48 it's not the last one, here comes page 2 15:10:57 Doh 15:11:10 yeah, I see 35 open items on my list 15:11:24 I need to get used to the SB layout. Was expecting something on the bottom of the list to indicate more. 15:11:24 * ttx clicks to see all 35 in one list 15:11:33 #link https://storyboard.openstack.org/#!/story/2001849 15:11:42 add information about ptls and liaisions to the releases repo 15:11:43 smcginnis: sounds like a good storyboard feature request 15:12:10 fungi: Not going to say I can implement it? :) 15:12:26 Have we fully switched to storyboard? 15:12:27 I don't understand this one 15:12:31 dhellmann: I think that should be rolled together with the "manifest" idea 15:12:32 what's the goal? 15:12:42 armstrong: For the releases repo we have. Not all projects though. 15:12:51 ttx: yeah, I think the stuff we talked about at the ptg replaces this 15:12:54 evrardjp: basically have a log of who was handling that release ? 15:12:56 evrardjp: Yeah, I added this, but now I can't recall the full context. 15:12:58 smcginnis: you can certainly add a story to openstack-infra/storyboard-webclient requesting it though 15:13:15 so we don't forget 15:13:22 I'm going to close this story as obsolete 15:13:28 yep 15:13:36 do we have another one for the mainfest thing? 15:13:36 although we did not create a story to track the master plan yet 15:13:43 Don't we need something similar for the changes we were dicussing in Denver? 15:13:45 so maybe we should rewrite this one 15:13:52 ++ 15:14:11 yeah because without the "master" plan, this is very obfuscated :p 15:14:12 the creation date would be misleading 15:14:50 so I'll close this one and someone will open another? 15:15:02 yes, I'll create the new one 15:15:04 Same basic concept though, just evolved over time. Not sure if the creation date would be an issue. 15:15:09 Either way works. 15:15:14 while you go down the list 15:15:30 #link https://storyboard.openstack.org/#!/story/2001848 New project checklist 15:15:54 I'll keep this one. There have been enough things pop up here and there for things that would be good to capture in something like this. 15:15:57 I think we still need this 15:16:21 #link https://storyboard.openstack.org/#!/story/2001846 Add verify library to process 15:16:33 after the ptg this one also feels obsolete 15:16:38 Not sure if this one is really an issue anymore 15:16:50 this one does mention automatically tagging so we could reuse it 15:17:17 Sure, let's hold on to this a bit yet. 15:17:41 #link https://storyboard.openstack.org/#!/story/2001845 Document policy on adding new deliverables to existing series 15:18:10 IIRC, the idea here was they can add to independent but not the series since these weren't actually a part of it at the time? 15:18:45 right 15:18:46 we kinda had that in OSA when a repo was renamed 15:19:19 Renames I think are fine. It's when something unofficial becomes official and they want to add pre-official releases I think. 15:19:20 so it's better be precise in the scope we are talking about here? 15:19:37 I'll add some notes so we remember. We should keep this one. 15:20:44 https://storyboard.openstack.org/#!/story/2003834 15:20:51 Stein release process evolution 15:21:12 Last task is actually one of the first to do... I miss task ordering! 15:21:43 Thanks 15:22:04 #link https://storyboard.openstack.org/#!/story/2001806 Support StoryBoard in release tooling 15:22:09 still on my to do list for sure, but if someone else is eager to do it instead i'm happy to hand it over 15:22:15 This one is actually to do with updating statuses. 15:22:22 that one has been on my plate forever 15:22:40 Still planning to do it if I ever manage to free up enough time 15:22:48 And doesn't seem like something we can break down into smaller steps to farm out. 15:22:58 Important to get to eventually. 15:23:00 it's a bit hard, due to the learning curve 15:23:10 so I would definitely keep it 15:23:20 will comment 15:23:32 #link https://storyboard.openstack.org/#!/story/2001802 Add commmand to setup for travis environment 15:23:36 i _think_ it's just a matter of importing python-storyboardclient and then mimicking what we do for lp 15:23:39 Another reno enhancement. 15:23:48 (wrt 2001806) 15:23:56 yeah, very low priority 15:23:57 oh, it;s assigned to you now! I know why I did not do it now 15:24:07 Sure, that's why. ;) 15:24:09 that would be used by folks outside of the openstack community 15:24:25 * dhellmann makes a note of ttx's task management approach 15:24:37 Useful for travis users for sure. 15:24:39 heh 15:24:51 s/useful/required/ 15:25:04 #link https://storyboard.openstack.org/#!/story/2001738 Support multiple --series arguments in list-deliverables 15:25:15 Seems like a low priority but useful thing too. 15:25:36 No reason the command can't be run multiple times though. 15:25:44 Any thoughts on this one? 15:26:23 seems like a good opportunity for someone to learn that code 15:26:26 let's keep it in case people look for easy ways in 15:26:28 and a very low priority 15:26:35 *cough* evrardjp *cough* 15:26:38 tagging low-hanging-fruit maybe ? 15:27:00 Tagged 15:27:25 #link https://storyboard.openstack.org/#!/story/2001692 Generate update emails for unreleased changes 15:27:42 smcginnis: :) 15:28:02 This one looks useful too. 15:28:23 Nothing critical by any means, but I think it would help raise awareness for teams of stable stuff sitting out there. 15:28:27 maybe we can get elod to help out there 15:28:46 Yeah, another good task to dig into things. 15:29:03 Not really a low-hanging-fruit though with some of the complexity of some of the steps involved. 15:29:11 But not too difficult either. 15:29:23 dhellmann: Leave it assigned to you for now? 15:29:39 I removed my name, but can pick it up if we need it done 15:29:58 the scripting is 90% done, so it's just the job and email sending that might be a challenge 15:30:18 OK, works for me. 15:30:32 #link https://storyboard.openstack.org/#!/story/2001690 Support release announcement emails to other mailing lists 15:30:36 Still needed? 15:31:03 Looks like maybe needs some input from fungi too. 15:31:14 checking 15:31:16 low prio but yes I would say 15:31:22 that was a thing I wanted for stevedore and some other libs 15:31:28 I think it's useful, but low priority 15:31:41 secrets are well documented on zuul ci docs nowadays 15:31:45 Yeah, for reno, cliff, stevedore, etc, could be really useful. 15:32:08 #link https://storyboard.openstack.org/#!/story/2001689 Ability to list a version on releases.o.o wihtout linking to artifact 15:32:23 smcginnis: what's the limit? Which libs will be announced or not? 15:32:47 why were we thinking we needed to authenticate smtp from those? 15:32:48 I think we need that one 15:32:56 evrardjp: I think just as needed when we find a deliverable that would be useful to announce to other communities. 15:33:03 we send to the release-announce list from job nodes currently without issue 15:33:08 fungi : because we would be sending email to lists not on lists.openstack.org? 15:33:16 maybe that's not really an issue then? 15:33:17 ahh, got it 15:33:20 fungi: Any issues sending to a ML on groups.io or something like that? 15:33:54 yeah, i have a feeling this is more a story about setting up an authenticated smarthost for jobs to send e-mail through 15:34:40 maybe we should only build that when we discover we need it 15:34:44 i'll add a comment now 15:34:50 fungi: Thanks 15:35:02 For listing on releases.o.o, looks like that was closed at one point. 15:35:08 Do we still want to get that? 15:35:18 I don't think it's really been an issue. 15:35:22 it does say "should the need arise" on building a smarthost for these 15:36:16 dhellmann: That's assigned to you right now. Thoughts? 15:37:12 I think we've been linking to things that don't exist? 15:37:25 or do we go through and remove those links? 15:37:35 I guess we cleaned up some at the end of rocky, didn't we? 15:38:05 Not sure honestly. 15:38:20 I'm OK with closing this and reopening it or making another one if we remember why we care about this 15:38:35 Looks like that happened once before. :) 15:38:37 heh 15:38:43 yeah, I had to turn comments on to see that 15:39:01 Maybe that's a sign that it's not really a big deal if we keep forgetting why we wanted it. 15:39:31 I guess the question is whether we care that we link to missing artifacts 15:39:43 Let's mark it invalid again. If it comes up again we can capture the full reasoning behind it. 15:39:45 and it seems like it hasn't been a huge issue 15:40:17 done 15:40:32 #link https://storyboard.openstack.org/#!/story/2001673 Announce job fails if diff-start does not exist 15:41:09 we probably ought to fix this one 15:41:11 Looks like a useful error handling thing to fix. 15:41:21 Updated the tag to stein. 15:41:37 #link https://storyboard.openstack.org/#!/story/1748164 Reno doesn't scale 15:41:40 ? 15:41:46 Ah, issues seen in nova. 15:41:56 they should just make fewer code changes ;-) 15:42:01 Hah 15:42:16 there are ways to configure the release notes build to not see the problem, and I think we did some of that 15:42:25 we could always work to make the code faster, of course 15:42:39 Looks like you added a useful workaround in the comments. 15:42:48 yeah, I think the task here is to document that better 15:43:02 No response, but come to think of it, does StoryBoard email notifications on comments? 15:43:09 it does 15:43:15 if you're subscribed 15:43:17 * dhellmann has many many comment emails from storyboard 15:43:18 I don't think I've ever received one, so no idea if anythign I've added has had questions or updates. 15:43:26 ah, I think you have to tell it you want to receive email 15:43:31 and if you don't disable story comments in your preferences 15:43:33 Wouldn't I be subscribed to stories I've added by default? 15:43:44 yeah, check your user prefs in the webclient 15:43:54 you have to "enable notification emails" globally too 15:43:59 https://storyboard.openstack.org/#!/profile/preferences 15:44:04 OK, I'll check that after the meeting. 15:44:14 ahh, right, we don't enable it by default 15:44:24 Hmm, seems like a questionable default. 15:44:36 Anyway, keep the reno scaling issue I guess. 15:44:38 we might consider adding it as a configurable default and changing it for new accounts or something 15:44:59 #link https://storyboard.openstack.org/#!/story/1748162 Tags at teh base of a stable branch get picked up twice 15:45:30 Looks like an issue when stable hasn't been tagged yet? 15:46:07 the tag at the base of a branch is on 2 branches 15:46:18 I think this is not a bug 15:46:45 Close it? 15:47:37 yeah, I just did 15:47:39 #link https://storyboard.openstack.org/#!/story/1748159 Empty configuration files cause exception 15:47:51 Looks like an opportunity to make reno a little more robust. 15:47:56 yeah, that's a real bug 15:48:36 #link https://storyboard.openstack.org/#!/story/1748158 earliest_version doesn't work with version from stable branch 15:49:03 dhellmann: Looks like you've provided some feedback there. Not a bug? 15:49:26 yeah 15:49:48 #link https://storyboard.openstack.org/#!/story/1734950 Support rst format not wrapped in yaml 15:49:52 For reno 15:50:04 Looks like a good low priority enhancement. 15:50:19 yeah 15:50:33 although now that I understand why someone asked for this I think it's probably a very very low priority 15:50:41 #link https://storyboard.openstack.org/#!/story/1728128 issue with repos that have checked in empty directories 15:50:50 I didn't realize git would allow this either. 15:51:14 Another low prio robustness opportunity 15:51:18 yeah, I think that's a relatively new thing 15:51:26 ++ 15:51:33 #link https://storyboard.openstack.org/#!/story/1712069 reno does not support git submodule 15:51:38 Good :D 15:51:40 wow, empty dirs in a git repo?!? 15:51:42 I bet all the others are the reno wishlist backlog 15:51:50 probably 15:51:57 Looking like it. 15:52:00 shoudl we switch to putting things on a board ? 15:52:15 I want to have an idea of must-have/should-have/good-to-have 15:52:32 and see if we have assignees for the first two 15:52:35 Yeah, we're running out of time again, so let's switch to that. 15:52:52 #link https://storyboard.openstack.org/#!/board/64 Task board 15:53:22 Isn't there a way that SB can automatically pull these into a board? 15:54:29 * ttx cleans up completed items 15:54:44 we could set up a board with query-based worklists but we left this one manual because you can't really mix automatic and manual 15:55:14 the board doesn't let you add items to a worklist if they are already in another worklist, so as soon as you have a list that shows the backlog automatically you have to set up the other lists using automated criteria, too 15:55:18 also do we really want a "done" list ? I know it provides sensory satisfaction but... 15:55:24 maybe I'll just clean it up 15:55:27 Can't we have To Do automatic for any openstack/release tasks, then manually move them to In Progress? 15:55:52 ttx: Yeah, I've always questioned Done items unless there's a follow-up reporting kind of need. 15:56:07 smcginnis: the idea was to select a subset of tasks for the Stein plan 15:56:19 Do All The Things 15:56:27 * ttx empties the list but keeps it 15:56:50 smcginnis : no, that's what I was trying to explain. If we set up a todo list as automatic, the board will not let you move items from that list to the other lists 15:57:07 That would be a good enhancement. 15:57:21 Well, we're almost out of time. Should we defer this and talk about the last topic quick? 15:57:22 it's a validation rule that tries to prevent you from adding something to todo and done manually, but it doesn't ignore the automatic lists 15:58:00 I'm adding what I have on my bucket 15:58:13 #topic Release changes 15:58:20 Since we only have a couple minutes... 15:58:29 dhellmann: one option would be to make story-based lanes which use specific tags in their queries and then control which lane(s) they appear in by changing the tags on them 15:58:31 #link https://etherpad.openstack.org/p/stein-relmgt-auto-release-change Proposed changes drafts 15:58:50 fungi: That's kind of what I was thinking. We have things tagged Stein. 15:58:54 fungi : yeah, it just seemed easier to manually add things to the todo list, then drag them into the other columns 15:58:55 Anyway 15:59:05 that way we have a backlog and a list of things we said we'd do this cycle 15:59:11 Please take a pass through the email drafts. 15:59:17 Should I send this early next week? 15:59:36 Though... I'm traveling Mon-Thur. 15:59:44 But I should have time to send it out. 15:59:57 one of us could send it if you would rather 16:00:04 but yeah, I think we should start talking about this asap 16:00:11 I agree. 16:00:17 At least email 1 16:00:27 I can send it, just want to make sure we are responsive if folks raise questions and concerns. 16:00:36 I'll be around next week 16:00:40 and Email 3 probably 16:00:43 Yeah, just the first email to start. 16:01:02 why wait for #2? 16:01:18 dhellmann: avoid input overload ? 16:01:23 Do we want to get feedback on #1 first before going too deep? 16:01:24 ok, that makes sense 16:01:28 maybe #2 the week after 16:01:35 sure 16:01:40 Over time. 16:01:54 cheers 16:01:58 If folks could take another look over the weekend and polish, I can send it early next week. 16:02:02 Thanks everyone! 16:02:10 thanks smcginnis! 16:02:14 #endmeeting