12:00:10 #startmeeting requirements 12:00:11 Meeting started Wed Aug 31 12:00:10 2016 UTC and is due to finish in 60 minutes. The chair is coolsvap. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:00:12 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:00:14 The meeting name has been set to 'requirements' 12:00:27 #topic Roll-call 12:00:37 o/ 12:01:07 o/ 12:01:12 prometheanfire dirk o/ 12:04:14 lets wait for a min before we start with meeting 12:05:25 lets move 12:05:27 #topic Announcements 12:05:37 I have a session accepted in barcelona with dims 12:05:42 #link https://www.openstack.org/summit/barcelona-2016/summit-schedule/events/15330/openstack-requirements-what-we-are-doing-what-to-expect-and-whats-next 12:06:56 #topic Any controversies in the Queue? 12:07:21 I don't think so 12:07:34 #link https://review.openstack.org/#/c/358470/ 12:07:37 the constant blocking of os-client-config sucks though 12:08:09 #link https://review.openstack.org/#/c/333264/ 12:08:15 #link https://review.openstack.org/#/c/363150/ 12:08:50 what's wrong with the gr-update? 12:09:25 "Updated from generate-constraints" review needs some attention for libraries which are only updated with this review 12:09:51 openstack libraries get updated with dims/dhellmann reviews 12:10:15 but others are not 12:10:32 right 12:10:54 I think we should strip the openstack libs out from that and get it merged today 12:11:09 prometheanfire: will you do that? 12:11:13 sure 12:11:45 #action prometheanfire to update the g-r review to strip the openstack libs 12:11:50 prometheanfire: thanks 12:11:53 that's just this review 12:12:10 not the actual job, but we should probably think about that 12:12:20 prometheanfire: yes we might need some work on the job as well 12:13:02 dhellmann: dims around? 12:13:15 pong coolsvap : what's up? 12:13:43 dims: reg. review https://review.openstack.org/#/c/358470/ 12:14:18 we are thinking if we can strip openstack libs in the "Updated from generate-constraints" reviews 12:14:41 dims: ^^ 12:15:18 prometheanfire: regarding the other two, I am not sure if we need a decision on the two new libraries before freeze 12:15:35 you mean the bot can propose 2 reviews? one with openstack libs and one without? 12:15:45 i'd run that by dhellmann for sure 12:16:20 dims: one the normal one by one updates, one outside deps 12:16:23 currently openstack reviews are merged with your/dhellmann/ttx patches generated after release change merge 12:16:39 openstack library updates i meant 12:17:01 but other libraries are getting ignored in the process 12:17:04 I updated the gr review for today 12:17:26 coolsvap : not all openstack libs will get reviews filed by me/dhellmann/ttx 12:17:32 coolsvap : example python-k8sclient 12:18:09 oh, should I add that back in? 12:18:17 dims: yeah thats another case 12:18:39 smaugeclient? 12:18:47 i'd rather have the bot differentiate into 2 buckets one from repositories owned by openstack and one outside 12:18:59 but remember it will make it harder to test 12:19:10 various combinations 12:19:19 so not sure what problem you are trying to solve here 12:19:22 my initial thought process was to merge "Updated from generate-constraints" reviews but we decided to merge individual reviews 12:19:34 but currently we are merging individual reviews 12:19:50 so the consolidated reviews are kinda ignored 12:20:06 which is fine, the bot does not get mad :) 12:20:12 either because it always fails with jenkins 12:20:29 but then we are losing on the updates for non-openstack libs 12:20:42 * dhellmann arrives late 12:21:03 coolsvap : so then the bot logging 2 reviews is a better idea i think 12:21:09 dims: yeah 12:21:33 well, added smaug/k8s 12:21:41 anyway, updated that review 12:21:44 oh, we shouldn't be ignoring the bot updates 12:22:00 dhellmann: yeah 12:22:01 dhellmann : ++ :) 12:22:21 it will be pretty complex to have the bot submit 2 reviews, I think. unless we build the changes in one repo, then figure out which lines to add to which patch? 12:22:23 I wasn't ignoring it, I tended to work on client updates first to shrink down be bot update, then try to get that merged once a week (on the weekend typically) 12:22:37 the idea is that all of the changes the bot proposes work together, so as soon as we start splitting them... 12:22:52 it might be safer to rebase the bot patch by hand to let it merge safely, and then approve it 12:23:18 dhellmann: the problem is that the bot submits so many changes at once that it almost always fails 12:23:44 that's what I've been doing (by hand edits to prune down the changeset) 12:23:47 ah 12:23:55 prometheanfire: yeah but its kinda ignored most of the times 12:24:00 well, maybe having it ignore our libs entirely would make it better? 12:24:03 afaics 12:24:05 coolsvap: not by me :P 12:24:15 though that might just lead to more failures 12:24:26 dhellmann: that's my early morning guess (before coffee) 12:25:00 yeah 12:25:40 lets monitor this for couple of weeks before making a decision 12:26:05 if its feasible to make it work by hand everytime 12:26:08 agreed 12:26:17 I just wanted to have people start thinking about it 12:26:26 * coolsvap too 12:26:43 can we just merge reviews that ignore the thigns that cause failures? 12:26:46 #note think about removing openstacklibs from gr updates (those that are already submitted seperately) 12:27:03 e.g. if externallibfoo-newrelease causes the bot to fail, letst just exclude that one to keep the proposal bot things mergeable 12:27:22 we just need to keep the backlog somewhere that investigates and fixes those issues as they happen 12:27:26 dirk: generally yes, last week was bad because there were a couple of updates that caused breakage 12:27:31 so it was multistep 12:27:48 yeah, figuring out which library is bad is non-trivial 12:28:22 lets move on 12:28:32 #topic Requirements freeze R-5 (Aug 29 - Sept 02) 12:28:32 agreed 12:28:46 #link https://releases.openstack.org/newton/schedule.html 12:29:00 tonyb sent the announcement on mailing list 12:29:06 #link http://lists.openstack.org/pipermail/openstack-dev/2016-August/102581.html 12:30:15 I have a question regarding that.. is that only about g-r changes (new additions, new base versions) or also about uc changes? 12:31:14 I think for both, dims dhellmann ^^ 12:31:17 both 12:31:20 afaik 12:32:04 * dims defers to dhellmann 12:32:18 so that means we should get an propsosal bot change merged pretty soon 12:32:21 yes, completely frozen except in the case of a bug in one of our libs 12:32:25 proposal.. 12:32:59 dirk: yeah thx dhellmann for the confirmation 12:33:13 lets monitor the change proposed by prometheanfire 12:34:13 coolsvap, dirk : I found it very helpful to create a process document for the release team in openstack/releases/PROCESS.rst. we could add details to that to make it clear about what is frozen when (if what I have there already isn't clear) 12:35:13 dhellmann: ack 12:35:40 #link https://github.com/openstack/releases/blob/master/PROCESS.rst 12:35:40 I think we have that down the agenda 12:35:49 #topic Coordinating with the release team 12:36:14 freezing requirements - http://git.openstack.org/cgit/openstack/releases/tree/PROCESS.rst#n107 12:36:35 stable branch for openstack/requirements - http://git.openstack.org/cgit/openstack/releases/tree/PROCESS.rst#n138 12:36:40 dhellmann: thanks for the pointer, I didn't see that yet 12:36:51 I think I saw in the requirements channel that tonyb said he would apply the -2 to all pending patches on his Thursday 12:37:11 or was it friday? 12:37:14 * dhellmann looks back 12:37:20 dhellmann: Friday 12:37:24 yeah, friday 12:37:43 which I think is late our thursday? how do timezones work again? 12:37:51 anyway, I think that's good 12:38:07 yes its late thursday your time morning for tonyb 12:38:14 he said his thursday, our friday 12:38:29 if we end up needing to land one of them that his -2 and he's away for the weekend we can start a new patch for the udpate 12:38:37 I may have that backwards 12:38:44 it's unlikely, but we never know so it's good to have a plan 12:39:04 dhellmann: I'll keep that in mind 12:39:06 prometheanfire : I checked the scrollback and he said friday 12:39:10 his friday 12:39:23 and since he lives in the future, that's our thursday 12:39:35 ah, right 12:40:04 * dhellmann hasn't had enough caffeine to do date math 12:40:33 okay lets convey this to tonyb if he's online tomorrow 12:40:49 #topic Barcelona Design Summit 12:41:10 tonyb has requested a (fishbowl) session for Barcelona 12:41:14 not going :( 12:41:37 maybe we can discuss the topics in some next meetings 12:42:05 we have a requirements session accepted in main track (not sure if everyone saw the announcements) 12:42:21 yep, saw that, good stuff 12:42:38 yarp 12:42:52 lets move on 12:42:54 #topic Mascot 12:43:12 The only mascot idea that had multiple votes was waterfall so tonyb asked Heidi to select that one 12:43:40 * coolsvap has no idea what was waterfall 12:44:19 I think its a good choice 12:44:22 I liked that one, it made me think "upstream" 12:45:16 cool 12:45:31 I think I proposed that one, it's our devel methodology right? 12:45:56 it fits for other reasons though 12:45:57 haha 12:46:51 #topic Tasks from Etherpad 12:46:58 I have to step away. Thanks everyone, and congratulations on becoming a big tent team! 12:47:11 thanks dhellmann 12:47:18 #link https://etherpad.openstack.org/p/requirements-tasks 12:47:29 any updates on this? 12:47:32 I've been keeping up on the other projects and failing requirements updates 12:47:34 * coolsvap has none 12:47:53 but no progress this week, one is stuck in legal, the other in review hell 12:48:10 that's all I have 12:48:39 thanks prometheanfire for the update 12:48:50 which one is stuck in review? 12:49:09 https://review.openstack.org/#/c/352501/ 12:49:13 #link https://review.openstack.org/#/c/352501/ 12:49:31 vitrage is the other one stuck in legal 12:49:46 it's the three under #20 in the list right now 12:50:07 ah okay 12:51:36 * coolsvap thought something stuck in requirements queue 12:51:46 moving on 12:51:50 #topic Volunteer for next meeting 12:52:01 I think we have volunteer for next two meetings 12:52:19 Sep 07 - tonyb 12:52:19 Sep 14 - prometheanfire 12:52:26 anyone up for sep 21? 12:53:29 alright lets see in next meeting again 12:53:47 I might be able to do sep 21, not sure yet 12:53:52 way too far in the future :) 12:54:08 dirk: alright we can always check in next meetings 12:54:24 skipping topic optional-requirements since did not see number80 12:54:36 #topic Open Discussion 12:54:57 anyone has anything to bring up in open discussion? 12:55:21 non 12:56:04 * dirk is good 12:56:26 okay back to #openstack-requirements. Thanks folks 12:56:31 #endmeeting