19:03:08 #startmeeting infra 19:03:09 Meeting started Tue Jul 26 19:03:08 2016 UTC and is due to finish in 60 minutes. The chair is fungi. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:03:11 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:03:13 The meeting name has been set to 'infra' 19:03:19 #link https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Agenda_for_next_meeting 19:03:24 #topic Announcements 19:03:24 o/ 19:03:26 I was interviewed for the Super User blog, talked about us some, hoping to raise visibility again of the fact that support for our team is essential to openstack: http://superuser.openstack.org/articles/why-infrastructure-is-the-key-to-openstack-s-big-tent 19:03:29 o/ 19:03:39 \o/ 19:03:52 pleia2: thanks for your tireless efforts marketing us! ;) 19:03:58 hehe 19:04:01 pleia2: thank you 19:04:08 pleia2: cool! 19:04:23 \o/ 19:04:23 i haven't read it yet, but it's probably way better than my superuser interview 19:04:24 Good stuff, thanks pleia2 19:04:24 nice :D 19:04:43 #info Reminder: late-cycle joint Infra/QA get together to be held September 19-21 (CW38) in at SAP offices in Walldorf, DE 19:04:45 #link https://wiki.openstack.org/wiki/Sprints/QAInfraNewtonSprint 19:04:54 we have started the xenial as default test platform switchover 19:05:02 oh, good one 19:05:11 had a minor issue getting job + branch selection correct for openstack-ansible which we fixed 19:05:14 no one else has screamed 19:05:24 i should start having people submit announcements to me (or add them to the agenda) ahead of time 19:05:27 so uh if you have time/interest feel free to help write chagnes to switch us over 19:05:31 clarkb: is that for stable releases as well, or are those in-time? 19:05:41 bkero: no stable remains on trusty so we roll forward on xenial 19:05:49 fungi: that sounds reasonable 19:05:59 So once newton is cut it will be xenial, then o, etc 19:06:03 yup 19:06:07 okay 19:06:09 we could discuss this (xenial testing) under the priority efforts if there's a lot of questions 19:06:29 #topic Actions from last meeting 19:06:35 #link http://eavesdrop.openstack.org/meetings/infra/2016/infra.2016-07-19-19.02.html 19:06:41 pleia2 Submit a spec to host an instance of limesurvey 19:06:43 i didn't see one in review... keep for next week? 19:06:52 it's half written, I'll finish it before friday (when I am traveling) 19:06:59 #action pleia2 Submit a spec to host an instance of limesurvey 19:07:07 right on 19:07:15 #topic Specs approval 19:07:24 none new this week 19:07:31 #topic Priority Efforts: A Task Tracker for OpenStack (Zara, SotK, anteaya, markus_z) 19:07:35 there was interest from markus_z in specifying a minimum warning period (maybe one month?) between tc acceptance and migration date 19:07:40 anybody think this is unreasonable to add to the spec? 19:08:06 I think its sensible 19:08:11 I do not think it is unreasonable 19:08:11 I'm fine with it :) 19:08:21 basically providing some stability for the community to finish up tool rewrites that were using the lp api so that they work with sb 19:08:36 without the api risking changing out from under them 19:08:46 makes sense 19:09:02 anteaya: are you up for writing that spec patch? 19:09:13 just adding a sentence will do 19:09:26 sure I will do so 19:09:46 we ideally would do some widespread announcement to the community once the tc accepts the migration plan anyway 19:09:53 oh yes 19:10:05 currently in talking to stakeholders 19:10:12 next step is tc 19:10:12 so just stating that there would be a minimum of one month from that until the actual migration or something 19:10:36 in reality i expect it will end up being well more than a month regardless 19:10:37 part of getting it on the tc agenda involves a post to the ml 19:10:52 I also think it will be well more than a month 19:11:06 but don't see any harm in stating a one month minimum 19:11:08 okay, anything else we need to cover on this effort in the meeting? 19:11:11 to ally fears 19:11:17 not from me 19:11:19 nor me 19:11:26 _o_ 19:11:56 clarkb: bkero: did you need me to squeeze the "Newton testing on Xenial" effort into today's agenda or were your questions basically addressed during the announcements? 19:13:02 I am good, but could use volunteers to help write teh changes likely 19:13:15 since I am traveling next week and much of the week after anddon't expect to have consistent internets 19:13:25 fungi: I'm good 19:13:35 so uh ping me if you can help 19:13:36 i expect those volunteers will come out of the woodwork anyway 19:13:38 clarkb: what changes are you referring to? 19:13:52 bkero: the changes to project-config that switch jobs to run on xenial 19:14:00 Got it, thanks. 19:14:01 #topic Priority Efforts: Newton testing on Xenial (clarkb) 19:14:27 #info Volunteers needed to draft project-config changes to switch more jobs to run on xenial 19:14:29 basically just need to update jjb and zuul togetehr to switch things, there are a few examples up and merged now 19:14:44 o/ I'll see if I can help out with that this week. 19:14:51 thanks clarkb, bkero! 19:14:54 because we decided to go explicit in the config its not a single 90 line python script we are modifying so lots of typing typing 19:14:56 bkero: thanks 19:15:12 #topic Infra mascot/logo (fungi) 19:15:14 typing typing typing and much seddings 19:15:18 #link https://etherpad.openstack.org/p/infra-mascot-ideas Infra mascot/logo ideas 19:15:23 looks like we have a bunch more now 19:15:36 are we all set for me to do a civs poll so we can rank preferences? 19:15:51 should i send ballots to the ptl electorate for infra? 19:15:58 or limit it to council members? 19:16:02 fungi: there is a deadline for tomorrow I think? so we better be ready? 19:16:16 yeah, i've talked to heidijoy about it 19:16:17 I would let everyone interested vote on it 19:16:22 on the voting does the top item win? 19:16:36 like if everything gets one vote and one thing gets 2 votes, does it win? 19:16:54 we need fallback choices in case of conflicts either within or (in some situations) outside the community 19:17:05 which is why you wanted to use civs which I think makes sense 19:17:14 so a condorcet ranking is a sane way to do that fairly, yes 19:17:19 yeah 19:17:48 i'll send ballots to our ptl electorate then and any of them who care can register their preferences 19:17:55 ++ 19:18:02 how long would make sense to leave it open? one week? 19:18:12 max one week 19:18:14 sounds good to me 19:18:15 yeaha 19:18:17 assuming thats fine with heidi that seems like a good time period 19:18:21 I think most folks will vote in 3 days 19:18:35 okay, i'll get that going right after the meeting, making it 6 days so i can close it before next week's meeting 19:18:42 yay 19:18:53 #action fungi start a poll for infra mascot 19:18:58 thanks fungi 19:19:18 #topic Contributor survey (pleia2) 19:19:25 #link https://etherpad.openstack.org/p/infra-newton-contrib-survey Contributor survey draft 19:19:31 not much new here, thanks to people who chimed in on the pad 19:19:44 next steps? 19:19:45 once we have a survey server up I'll put the questions in it 19:19:54 ahh, right-o 19:19:58 we can likely chat about it again before we announce it 19:20:13 sounds good. can come off the agenda for now i guess? 19:20:17 yes, thanks 19:20:19 and you'll readd it once we're close 19:20:25 perfect 19:20:25 * pleia2 nods 19:20:35 pabelanger: are you back yet? 19:21:04 i'll reorder these two topics to give pabelanger a chance to get back to his computer 19:21:12 #topic Gerrit 2.11.4 updates (zaro) 19:21:29 looks like we have several fixes merged, a couple still under review 19:21:30 so it's been a while since we updated gerrit 19:21:46 i thought i would see what we are interested in updating. 19:21:49 yep, as we discussed a few weeks ago, this seems like a good idea 19:22:07 what would be involved in updating? 19:22:11 the gson update is needed for storyboard plugin 19:22:15 :D 19:22:33 anteaya: it means reving the version of gerrit we have on review.o.o 19:22:36 zaro: what's the deal with maven and bouncy castle? 19:22:39 #link https://review.openstack.org/345540 re=nofollow spam deterrent (merged) 19:22:42 it would involve a patch to make that update 19:22:56 #link https://review.openstack.org/260324 puppet code highlighting (merged) 19:22:57 zaro: right, so how much downtime, do we need db backups, what steps are we following? 19:23:19 #link https://review.openstack.org/311903 fix bouncy castle lib download (merged) 19:23:22 or just restart gerrit once patches are merged? 19:23:26 (I am watching the its-storyboard patches like an oversized hawk; thanks again for working on those) 19:23:28 yeah that one :) 19:23:49 corvus: the libs disappeared from the repo so we needed to update the link in gerrit. i don't know why it disappeared though 19:23:57 #link https://review.openstack.org/340157 update gson lib, required for its-storyboard plugin (open) 19:23:58 did they update that upstream too? 19:24:04 anteaya: no reindex required 19:24:11 zaro: thank you 19:24:14 #link https://review.openstack.org/#/q/status:open+project:openstack-infra/gerrit+branch:openstack/2.11.4 other open changes for 2.11.4 19:24:22 zaro: or does upstream use a different version of bouncy castle...? 19:25:11 yes, upstream was fixed the same way 19:25:51 ok cool 19:25:59 i mean, weird, but cool. :) 19:26:11 the first 3 changes on the infra meeting page is required, so i'm wondering about the others open changes for 2.11.4. 19:26:34 zaro: the first three changes have all merged 19:26:42 i would like to see https://review.openstack.org/267927 go into the next build too, as that makes our election process more transparent 19:26:51 opps i mean first 4 changes. including the open gson lib update change 19:26:54 #link https://review.openstack.org/267927 Remove required access permission to list account emails 19:27:08 fungi: that would be awesome 19:27:14 zaro: ah okay 19:27:25 i think we also wanted https://review.openstack.org/#/c/267927/ or maybe at least anteaya wanted it 19:27:34 i believe that's the only remaining blocker for the revised electoral roll script to be runnable by non-gerrit-admins 19:28:07 are we okay with that from a privacy perspective? 19:28:24 zaro: yes, I fungi linked it above, and yes it would be awesome to have in 19:28:31 s/I// 19:28:37 ohh didn't fungi want this one too? https://review.openstack.org/#/c/277538/ 19:28:56 i guess most emails are already publically accessible -- either as the primary account email (is displayed on most pages in the ui) or as emails to match git authors (displayed on git commits) 19:29:04 i can't think of any other cases... 19:29:06 corvus: so gerrit already makes your preferred e-mail address queryable 19:29:09 yes 19:29:11 that 19:29:24 ++ 19:29:28 this just extends it to the additional non-preferred addresses too 19:29:28 sounds good then 19:29:29 that's what I was thinking 19:29:56 zaro: can we get a consistent topic on all the patches you want to see merged prior to the update? 19:30:02 zaro: oh! right, 277538 while not necessary, will simplify the electoral roll generation script significantly 19:30:15 takes out a lot of the current guessing around figuring out when a change actually merged 19:30:21 well ok, i conclude that we want 340157, 267927 and 277538 merged? 19:30:39 zaro: sounds fine to me 19:30:42 #link https://review.openstack.org/277538 Append submitted info to the change REST endpoint. 19:31:10 zaro: those three sound great, yes 19:31:37 anybody have a vested interest in " 19:31:39 er 19:31:46 ok, infra-core do your thing and i'll prep a patch for the update 19:31:55 anybody have a vested interest in "Increase default HTTP incoming connection queue" or "Init plugins' AllRequestFilters, even if they are loaded after startup"? 19:32:37 ohh, maybe this one too? https://review.openstack.org/#/c/268689/ fix javamelody plugin updates 19:33:19 is that a significant bump in javamelody featureset over what we have now? 19:33:50 no changes to javamelody itself it just fixes a problem with updating the javamelody plugin while gerrit is running 19:34:37 it's a fix to the mechanism of deploying the plugin. 19:35:30 also i can prep a change to update on review-dev.o.o and test it out there first before we update review.o.o 19:35:31 was it causing a problem for us, or is it likely to do so if we don't approve that backport? 19:35:50 just wondering what the impetus was for backporting that 19:36:08 currently if we attempt to update the javamelody plugin without restarting gerrit it won't work. 19:36:22 i can't remember having done that 19:36:34 I thought the plugin version was tied to the war build anyways 19:36:41 ok, then it's probably not important. we can just restart gerrit when updating javamelody 19:37:18 to clarkb's point, i think our current process for updating gerrit plugins has an implicit gerrit restart baked into its assumptions 19:37:31 clarkb: it is build against our fork of gerrit core but it's not packaged with the war 19:37:54 so yeah, this seems like something we can probably drop. we'll consume the fix the next time we upgrade to a newer gerrit release containing it 19:37:57 ok, then we should skip it. 19:38:18 i think it only failing for javamelody anyways. 19:38:46 thanks. anything else you wanted to cover on gerrit upgrades in the meeting? we should likely get the outstanding changes we've identified reviewed/merged prior to scheduling an upgrade maintenance 19:39:29 so action is to merge 340157, 267927 and 277538 ? 19:39:35 yep 19:39:48 cool, that's it. 19:39:59 #info We want 340157, 267927 and 277538 merged before we schedule a production Gerrit upgrade maintenance 19:40:08 ohh do we need to identify a date? 19:40:12 #undo 19:40:13 Removing item from minutes: 19:40:16 #agreed We want 340157, 267927 and 277538 merged before we schedule a production Gerrit upgrade maintenance 19:40:26 wanted to make that an agreement, not an info 19:41:10 zaro: we can probably leave scheduling for after those changes merge. it ought to be a relatively quick outage so lengthy warning announcement is probably not warranted. a few days heads up should be sufficient 19:41:22 ++ 19:41:47 pabelanger: are you back at your keyboard yet? 19:42:45 #topic Release git-review (fungi) 19:42:56 giving pabelanger a few more minutes to get back 19:43:10 i want to tag a new git-review release later this week 19:43:37 will that come with voting ? i.e. git review plus wip'it ? 19:43:42 anybody feel like reviewing probably non-impacting open changes for any we should squeeze in? 19:43:48 i kind of remember a change for doing that in review 19:43:51 fungi: do you think its worth adding a note about gerrit without bouncy castle and newer git support in git-review? 19:43:54 #link https://review.openstack.org/#/q/status:open+project:openstack-infra/git-review open git-review changes 19:43:54 er newer ssh support 19:44:13 I don't actually know how supporting gerrit without bouncy castle is considered 19:44:14 what's included in that git-review release? 19:44:21 clarkb: to the readme/manpage? probably wouldn't hurt 19:45:38 fungi: ok I can try to write that up quickly since I know at least one other person/group was affected 19:45:48 rcarrillocruz: #link https://review.openstack.org/#/c/281504/ 19:45:55 zaro: see stuff above the 1.25.0 tag for what's already landed in master 19:46:01 #link https://git.openstack.org/cgit/openstack-infra/git-review/log/ git-review log 19:46:23 oh, there you go, thx 19:47:40 rcarrillocruz: we want to avoid baking openstack community gerrit configuration assumptions into git-review features, so i don't expect a "wip" feature in git-review until we have a generic way of setting wip status in gerrit 19:48:24 fungi: i suggest this one, #link https://review.openstack.org/#/c/285620/ 19:48:29 a compromise has been floated to add a feature for submitting an arbitrary review comment/vote immediately following a patchset push 19:48:31 yeah, makes sense 19:48:46 fungi: add a --wip=Workflow-1 into .gitreview 19:49:34 not a bad idea 19:49:36 fungi: i don't understand your stament 19:49:59 is the 'wip' vote the issue? 19:50:16 that feature is generic to set any vote, not specific to wip 19:50:28 zaro: proposing a git-review feature where you can ask it to submit a review comment with vote (e.g. workflow -1) when pushing a patchset 19:51:00 that feature is not specific to wip vote. you can set Code-Review vote or verified vote or any vote 19:51:18 you can do that with gerrit itself 19:52:04 zaro: yes, this comes on the heels of prior patches proposing "wip" features to git-review where our advice was to implement what 281504 is doing 19:52:55 zaro: i was probably not being clear. 281504 makes sense, and it's a good compromise 19:53:38 * rcarrillocruz went straight to the commit msg example and didn't realize the change is about setting whatever vote 19:53:44 so ++ 19:54:08 we can change the example to Verified vote to ease your concern 19:54:12 pabelanger: back yet? 19:54:38 heh 19:54:42 no need :P 19:54:48 zaro: i don't have any concern. i'm in favor of the idea of 281504 i just haven't reviewed the implementation yet. it's on my short list now 19:55:32 sounds good, and i think i've already reviewed just about all of the oustand changes. 19:55:33 #topic Open discussion 19:55:44 thanks zaro, it's a huge help 19:55:54 Just a heads up I am likely to be AFK for all of next week and much of the week after. Visiting family and they live without Internets 19:56:09 I'd like to see a lesser puppet run for simpler image builds 19:56:14 clarkb: yay a break for you 19:56:17 clarkb: have fun 19:56:40 prometheanfire: can you elaborate? 19:56:46 clarkb: those are the family i avoid visiting :) 19:56:51 adding gentoo support to all the puppet modules is becoming a real pain 19:57:10 prometheanfire: we're trying to unpuppetize our image builds over time 19:57:44 I realize that, and that'd likely help 19:57:54 this is mostly an artifact of the bad old days when we only had static job nodes and they were built the same way as our other servers 19:57:55 zaro: I have no idea how they survive (well really I do they live on a mountain side in hawaii and don't need Internets) 19:57:58 it the thing I've been stumbling on 19:58:45 I'm also AFK next week, conference+giving a tutorial in Mumbai and then may try to dodge rain as I attempt to tourist a bit during monsoon season 19:58:57 :S 19:59:15 pleia2: yay, break for you too 19:59:38 clarkb: well, i would compromise for hawaii though :) 19:59:40 pleia2: clarkb: don't get washed away! 19:59:44 we need you back soon 19:59:46 clarkb: aren't the mountains in hawaii volcanoes? :) 19:59:47 all that puppet needs to do with the gentoo image is create a user 19:59:55 mtreinish: they are and its active, all good fun 19:59:58 okay, that's it for times 19:59:59 so adding support for everything else seems excessive 20:00:04 #endmeeting