14:31:40 #startmeeting releaseteam 14:31:40 Meeting started Fri Jun 3 14:31:40 2016 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:31:42 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:31:44 The meeting name has been set to 'releaseteam' 14:32:10 Agenda is on R-18 on https://etherpad.openstack.org/p/newton-relmgt-tracking 14:32:35 dims: around ? 14:33:29 * dims wanders back :) 14:34:02 So.. we have a number of newton-1 tag requests 14:34:13 Which are all blocked for various reasons 14:34:35 Manila-UI https://review.openstack.org/#/c/324788/ I think we can process if you agree 14:34:36 right, we need to recheck some of the red ones 14:35:02 +2 ttx 14:35:18 OK, let me process it now 14:35:19 also give me a heads up once they're all flushed for the day. infra will defer the gerrit maintenance window until they all get through 14:35:32 list-changes looks good for that one 14:35:50 hmm, that said it's not a pure newton-1 tag so could wait 14:36:24 ack. 14:36:58 ttx : we could do the ironic ones 14:37:04 from yesterday 14:37:16 dims: I'm only discussing newton-1, since we are Friday 14:37:44 gotcha. 14:38:10 y i don't see anything else ready 14:39:06 We may do Kolla but prefer to wait for dhellmann 14:40:33 dhellmann: welcome, we were discussing the last newton-1 tag requests. I compiled them on https://etherpad.openstack.org/p/newton-relmgt-tracking R-18 14:41:13 * dhellmann reads scrollback 14:41:44 Astara needs a bypass 14:42:07 Kolla - I think we can process it but would rather have your +2 on that version number 14:42:14 https://review.openstack.org/#/c/325068/ 14:42:18 ttx : of the 3 kolla reviews, https://review.openstack.org/#/c/325068/ is the only one for n-1 14:42:25 yep 14:43:46 ttx: what's wrong with the kolla version? 14:44:08 Just wanted agreement that this meant they will follow milestones all cycle except for final 14:44:17 since they are cycle-trailing 14:44:44 hmm, yeah, I made that assumption but we could verify with sdake before processing it 14:45:02 I think we can process it tbh 14:45:18 The Kolla-K8s one is weird 14:45:39 It's independently-released, I don't think we support b1 tags there 14:45:42 processing kolla wfm 14:45:44 yeah, we don't 14:45:48 I'll comment 14:45:50 ok 14:46:06 that's also a weird beta of a non-1.0 release 14:46:11 The last 4 couldn't be processed due to gate issues, but also some were submitted only a few hours ago 14:46:26 hmm, that probably should be cycle-trailing too eventually 14:46:29 I think anything submitted "today" is late and we should reject them. Thoughts? 14:47:18 Neutron one, we need to discuss which deliverable model to follow, between teh one being proposed in governance and the one currently in 14:47:33 dhellmann: given state of gate I'm fine with cutting some slack but close the door now 14:47:49 ok, I guess that's fair 14:47:50 dhellmann : guess we need to check who else has not even bothered to file a review 14:48:08 dims: next topic 14:48:23 dhellmann: so, what should we follow for neutron ? 14:48:51 separate deliverables for aas like suggested in https://review.openstack.org/#/c/323522/ or all-in-one like requested in https://review.openstack.org/#/c/323776/ 14:49:05 I guess that's an issue with asking deliverables/release models frozen by n1 14:49:18 when they change precisely at n1 :) 14:49:20 if they wanted to change this they should have done it last week 14:49:35 yeah, that seemed to be why armax was in a hurry to get the governance change approved 14:49:39 ok, so we should process the tag and deny the tag change 14:49:53 deny the deliverable map change 14:49:56 I mean 14:49:57 thinking... 14:50:37 dhellmann: it's difficult to go back, because the change says they are cycle-with-intermediary 14:50:38 I don't like the release model tag change. If those projects are not seeing a lot of action, but are still immature, that's not a reason to use the intermediary model, which we intended for more mature projects that wanted to release frequently to have. 14:51:00 ok, so process request and then say too late to change the tags/deliverable map 14:51:22 deliverable could still be modified I guess 14:51:28 but not release model 14:51:33 I'm fine with changing the deliverable mapping if we don't see any side-effects for defcore 14:51:40 i (or someone on the vmt) is still likely to repropose a similar change to split them into separate deliverables if only so we can have vulnerability:managed not apply to unintended repos 14:52:04 ack 14:52:11 fungi : ++ 14:52:12 OK, I think we can step to next topic 14:52:30 so what's the decision? 14:52:47 dims: we process https://review.openstack.org/#/c/323776/ as-is 14:52:56 cool 14:53:00 and we'll -1 the release model change on https://review.openstack.org/#/c/323522/ 14:53:20 #topic release model changes 14:53:22 yeah, I've been trying to indicate that my -1 is for the release model change there already but I'm not sure I'm being clear 14:53:30 We have two other late release model change requests 14:53:35 PuppetOpenStack https://review.openstack.org/#/c/323027/ 14:53:39 Fuel https://review.openstack.org/#/c/306071/ 14:53:49 i'll ping the fuel guys 14:53:51 o/ 14:54:03 The Fuel one was posted a long time ago but the proposer never rebased it 14:54:19 I'm fine fast-tracking it if we can get the rebase and the PTL +1 now 14:54:56 The PuppetOpenStack one is stuck in the one-week-maturation-for-governance-changes hell 14:55:18 I don't expect people to object to it, but we are still supposed to wait one full week 14:55:29 cool, thanks! 14:55:30 I'm fine with considering it meeting the deadline though 14:55:37 I agree 14:56:25 OK, I noted to decisions on the etherpad 14:57:09 #topic newton-1 postmortem 14:57:12 I'll work on the astara validation bypass monday, I don't think there's a rush there 14:57:27 So, who missed the boat ? 14:57:33 * ttx looks 14:57:55 some of the networking projects 14:58:11 dhellmann : ttx : am getting on a call in a few mins. 14:58:19 dims : ack 14:58:28 barbican 14:58:52 training-labs (?) 14:59:35 murano-agent maybe 14:59:44 why is training-labs milestone-based I wonder 14:59:51 murano-apps, murano-dashboard maybe 15:00:06 networking-ovn 15:00:17 * russellb holds head in shame 15:00:39 aodh ? 15:00:40 it looks like murano is in https://review.openstack.org/#/c/325244/ but is missing apps 15:00:48 ceilometer 15:00:55 trove 15:01:07 trove-dashboard 15:01:16 trove-image-builder 15:02:44 yeah, that's a bit of a longer list than I expected 15:02:49 I'm glad this is only the first milestone :-) 15:02:57 murano is actually only missing murano-apps 15:03:54 * amrith tries to follow references to trove 15:04:01 * redrobot pokes head in 15:04:19 amrith : we're discussing projects that missed the milestone deadline 15:04:45 dhellmann: I think I caught them all 15:05:01 dhellmann: probably a small report is in order, once you have processed the backlog 15:05:03 ttx: that looks good, I'll double check that list when I copy it over to the tracking etherpad for later reference 15:05:06 yeah 15:05:20 I have a call I need to hop on, so I'll catch up with those things after that 15:05:28 on the process issues side... 15:05:34 I wasn't planning a public email list, unless you think that's necessary? 15:05:36 oh, you need to jump ? 15:05:46 now ? 15:05:51 I can take a minute or two 15:05:57 for those that missed, i'm wondering what your preferred action is 15:05:59 ok, let's go quick 15:06:04 submit late? skip it but do better for n-2? 15:06:14 dhellmann: slow check queue you already have a change about 15:06:18 russellb: skip 15:06:27 russellb : we won't take the late tag, so just keep an eye on the deadline for n2 15:06:27 ack thanks 15:06:45 ttx: jeblair had an alternate suggestion of not requiring a patch to pass check before entering our gate 15:06:47 dhellmann: we also have projects for which I wonder if there is not confusion as to what they are supposed to do at n-1 15:06:57 they don't seem to like the idea of having a different priority check queue 15:06:58 ttx, I'm in that camp 15:07:20 like projects that do intermediary releases but still posted 'n-1' releases 15:07:37 I listed 3 in the meeting etherpad 15:07:46 the first milestone I have listed is $-16 for spec freeze. we weren't releasing anything at n-1 15:07:48 yeah, I noticed a few of those. I think it's fine, we just won't prioritize them over the milestone projects 15:07:53 dhellmann: i think it's more an issue of added complexity to have more than one voting check.* pipeline 15:08:14 amrith : trove is listed as following the cycle-with-milestone release model 15:08:19 dhellmann: ok, otherwise let's discuss n2 tasks next week 15:08:29 fungi : ah, ok, that's fair too -- either way, we started the discussion of how to solve the problem 15:08:30 #topic Other 15:08:45 dhellmann: a few pointers you should look at 15:08:48 dhellmann: also if there are changes in stuck waiting in check which are impacting release deadlines, i consuder that plenty justification for an infra-root admin to jump in and enqueue them directly into teh gate for you 15:08:52 Julien's question about release jobs @ http://lists.openstack.org/pipermail/openstack-dev/2016-June/096475.html 15:08:57 Project team guide release process updates: https://review.openstack.org/#/c/324540/ 15:09:12 (and more generally all release-orienetd people) 15:09:14 wow, my typing just got really, really terrible 15:09:18 Release ACL dance infra-spec https://review.openstack.org/#/c/322863/ 15:09:43 I will likely skip my release day next Wednesday. Will try to do things but =can't guarantee 15:09:51 dhellmann: that is all 15:09:59 dhellmann, I was following http://governance.openstack.org/reference/tags/release_cycle-with-milestones.html and had listed the points when we were going to release things. I put them into the release schedule. http://releases.openstack.org/newton/schedule.html 15:10:12 ttx: I'll look for that thread and reply. I think we want a new job template for jd__'s case 15:10:33 amrith : if you pick the milestone model, you don't get to make up your own schedule 15:11:00 just a quick update on the release automation work, we do now have a newton cycle signing key published on the keyserver network (and i've signed it with my personal key as well) 15:11:03 ttx: for the other things I'll make sure I've voted on the latest version of the patch 15:11:05 #link http://p80.pool.sks-keyservers.net/pks/lookup?op=vindex&search=0x64DBB05ACC5E7C28&fingerprint=on&exact=on 15:11:14 fungi : \o/ 15:11:25 i'm currently hacking on the automation to handle our key updates from hiera when we rotate them each cycle 15:11:28 dhellmann, ok. I'll push a release for trove n-1. and make sure I do n-2 as well 15:11:37 amrith : it's too late for n-1, so don't worry about that one 15:11:48 ok, n-2 it is. 15:11:49 since it's basically the automation we also need to install that key on the signing worker 15:11:53 will make sure I don't miss it. 15:12:02 fungi : cool, sounds like good progress 15:12:19 dhellmann: less progres sthan i'd hoped for. week got crazy (as it always seems to) 15:12:21 ok, I need to drop off, is there anything else critical? 15:12:29 fungi : still, moving forward :-) 15:12:38 dhellmann: nope nothing critical 15:12:47 ok, thanks, and sorry about the screwy schedule today 15:12:59 np 15:13:01 #endmeeting