19:01:16 <fungi> #startmeeting infra 19:01:17 <openstack> Meeting started Tue Oct 6 19:01:16 2015 UTC and is due to finish in 60 minutes. The chair is fungi. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:01:20 <openstack> The meeting name has been set to 'infra' 19:01:22 <rfolco> o/ 19:01:24 <dougwig> o/ 19:01:25 <ruagair> o/ 19:01:30 <fungi> #link https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Agenda_for_next_meeting 19:01:33 <zaro> o/ 19:01:38 <mmmpork> o/ 19:01:42 <krtaylor> o/ 19:01:48 <fungi> #topic Announcements: Stackforge namespace retirement 19:02:04 <fungi> #link https://wiki.openstack.org/wiki/Stackforge_Namespace_Retirement 19:02:23 <jeblair> (i still need to send a follup mail; hoping to do that today) 19:02:24 <fungi> just a reminder, this is still scheduled for october 17 19:02:30 <fungi> thanks jeblair 19:03:01 <fungi> #action jeblair send followup reminder to lagging stackforge projects about namespace retirement 19:03:10 <anteaya> did we ever set a time on oct 17? 19:03:19 <anteaya> or shall we discuss that later? 19:03:29 <fungi> we can discuss it now i suppose 19:03:29 * anteaya notes we are still in announcemnents 19:03:37 <fungi> if we keep it brief 19:03:47 <anteaya> anyone have any thoughts about what time gerrit should go offline? 19:03:57 <fungi> 18:00 utc? 19:04:01 <anteaya> wfm 19:04:08 <fungi> who expects to be working on it? 19:04:11 <anteaya> me 19:04:14 <mmmpork> how is the project name change going to be run? 19:04:22 <fungi> we probably want to leave ourselves plenty of time given the scale of the change 19:04:23 <jeblair> me, and 1800 works, so does earlier 19:04:39 <anteaya> fungi: I'm guessing about 2 hours downtime anyway 19:04:40 <jeblair> (but we also might need extended prep before, so maybe not schedule the outage too early) 19:04:41 <fungi> i didn't want to pick something too early for west-coasters 19:04:43 <anteaya> is that too much? 19:05:28 <mordred> o/ 19:05:29 <fungi> though if SergeyLukjanov, yolanda or jhesketh are wanting to pitch in on the maintenance we can consider less americentric windows 19:05:34 <mmmpork> also have we defined a verification process for checking the change result before it goes live? 19:05:40 <mordred> (on a plane, landing in 30 minutes, fwiw) 19:06:00 <pleia2> I'll be on a flight to japan during the rename, so one less west-coaster to worry about 19:06:06 <jeblair> mordred: are you indicating you are here, or that you will be around on oct 17? :) 19:06:27 <anteaya> pleia2: thanks for letting us know, safe travels 19:06:29 <jeblair> pleia2: that sounds fun, we should all go 19:06:38 <pleia2> jeblair: yeah, let's do it 19:06:48 <fungi> for those without a calendar (or the `cal` utility) handy, keep in mind this is one week from saturday 19:07:04 <yolanda> fungi, i cannot be here on 17th october, i have some family appointments. But thanks for the consideration, and i'll try to help in other future tasks 19:07:25 <anteaya> fungi: looks like it is you me and jeblair so far 19:07:27 <fungi> yolanda: no problem. i'm sure we have lots of people on hand, and thanks for the heads up 19:07:31 <anteaya> I'll do as much prep as I can 19:07:40 <mmmpork> i'll be around for it 19:07:50 <mmmpork> but i'd like some answers to the questions i just asked :) 19:08:13 <clarkb> I will be around 19:08:14 <fungi> okay, to rephrase. any objections to 18:00 utc (or think there's a better time)? 19:08:22 <anteaya> no objections 19:08:23 <clarkb> 1800 wfm 19:09:00 <mmmpork> no objections here 19:09:07 <fungi> #agreed Stackforge namespace retirement maintenance will commence at 18:00 UTC, Saturday, October 17 19:09:07 <vmbrasseur> mmmpork Those are great questions. 19:09:36 <fungi> mmmpork: non-scheduling-related questions can be handled later in the meeting or in #-infra 19:09:37 * ruagair is around too 19:09:44 <mmmpork> fungi: ok thanks :) 19:09:53 <fungi> #topic Actions from last meeting 19:10:03 <fungi> #link http://eavesdrop.openstack.org/meetings/infra/2015/infra.2015-09-29-19.01.html 19:10:13 <fungi> Clint: write script to prepare layout.yaml change for migration 19:10:14 <Clint> #link https://review.openstack.org/#/c/228002/ 19:10:49 <Clint> and as sample "output" for a subset of names and using an earlier patchset, there's 19:10:52 <Clint> #link https://review.openstack.org/#/c/229601/ 19:10:57 <mordred> jeblair: sorry, that was an indication of my skills at attneding thi smeeting. I will be here on the 17th 19:11:14 <fungi> cool, thanks Clint! (and anteaya and mmmpork who seem to be the primary reviewers so far) 19:11:18 <jeblair> mordred: woot :) 19:11:21 <anteaya> that output used patchset 12 of the script 19:11:32 <jeblair> does that script cover all the necessary files now? 19:11:36 <mmmpork> i've added some unit test coverage so we can check the file names we are changing and the data we'll be writing out 19:11:38 <jeblair> (i mean, as far as we know :) 19:11:43 <Clint> as far as i know 19:11:43 <mmmpork> jeblair: idk does it? ;) 19:11:45 <anteaya> jeblair: patchset 12 looked fine to me 19:11:56 <anteaya> jeblair: as far as I know 19:12:04 <mmmpork> there's a test that i wrote that lists the files we're changing 19:12:10 <jeblair> mmmpork: exactly :) rephrased: i take it there are no known omissions at this point? 19:12:19 <mmmpork> if someone could take the time to review that and make sure our expectation are covered there, that would be great 19:12:45 <Clint> the only known issue at this time is that some comments will need to be replaced by hand 19:12:58 <anteaya> Clint: in gerrbot/channels.yaml 19:13:02 <Clint> right 19:13:03 <mmmpork> https://review.openstack.org/#/c/228002/27/tools/tests/mass_rename_projects_unit_test.py 19:13:06 <fungi> the sample change looks good on a brief pass. was there a plan to handle the sql update queries? 19:13:09 <anteaya> as noted in the commit message in the script 19:13:23 <Clint> fungi: no, i thought that might be done by the ansible stuff, which i haven't seen 19:13:36 <fungi> jasondotstar: that's what you were hacking on? 19:13:39 <anteaya> fungi: as far as I know the sql queries were not part of this script 19:14:00 <Clint> i was thinking this script would not extend outside the boundaries of the project-config repo 19:14:20 <anteaya> Clint: well so far it doesn't 19:14:29 <Clint> right 19:14:34 <mmmpork> do we have a checklist somewhere of all the pieces that need to be updated? 19:14:42 <anteaya> so many steps on of the workflow are not automated 19:15:00 <Clint> mmmpork: there's https://etherpad.openstack.org/p/project-renames-Septemeber-2015 19:15:11 <fungi> oh! right, i skimmed and misread the git mv commands comment you added 19:15:12 <jeblair> Clint, mmmpork: and http://docs.openstack.org/infra/system-config/gerrit.html#renaming-a-project 19:15:23 <anteaya> fungi: yes I did to at first 19:15:32 <anteaya> fungi: thos commands are for the acl file path renames 19:15:54 <anteaya> which Clint will run by hand before submitted the patch to be merged 19:16:11 <jeblair> if we don't have an ansible thing, we have fungi's "script that outputs the sql" we could use 19:16:15 <fungi> so, anyway, if worse come to worse we have that already scripted (more or less) 19:16:21 <fungi> #link https://review.openstack.org/222726 19:16:28 <anteaya> ast you can see in the sample output 19:16:36 <anteaya> #link script sample output https://review.openstack.org/#/c/229601/ 19:16:55 <mmmpork> so i'm seeing a bunch of things in a bunch of different places for this project 19:16:55 <fungi> okay, seems like we're on track then 19:16:56 <anteaya> jeblair: yeah I haven't seen one yet and haven't looked at sample output 19:17:10 <anteaya> fungi: I feel comfortable with where the patch should be 19:17:14 <mmmpork> and it's still not clear how we're going to do a test run and verify before pushing this change 19:17:23 <anteaya> fungi: you okay to create the working etherpad for this? 19:17:59 <fungi> anteaya: yeah, though for sanity it may make sense to not paste the filesystem and database updates into the etherpad directly 19:18:07 <anteaya> agreed 19:18:21 <anteaya> I'm open to however you want to address running those steps 19:18:35 <fungi> #action fungi start maintenance plan etherpad for stackforge retirement 19:18:38 <anteaya> as long as we can have the steps noted for when they are done 19:18:42 <anteaya> thanks 19:19:00 <fungi> we had one other action item from the last meeting... 19:19:07 <fungi> jhesketh: look into why the swift upload indices are still not appearing 19:19:34 <fungi> jhesketh may not be around 19:19:43 <fungi> anybody happen to know the status on this? 19:19:48 <anteaya> fungi: do you have a new cpu fan for your workstation? 19:20:06 * anteaya does not know the status of swift upload indices 19:20:11 <fungi> anteaya: nope, i'll have contingencies in place this time 19:20:16 <anteaya> fungi: wonderful 19:20:28 <mordred> fungi: we've started our descent - sorry, hav eto drop 19:20:30 <Clint> btw, the project-config script takes a different input format from the one in https://review.openstack.org/222726 19:20:34 <anteaya> mordred: safe landing 19:20:48 <clarkb> fungi: its working now 19:20:58 <anteaya> clarkb: yay 19:20:59 <fungi> clarkb: excellent 19:21:11 <clarkb> however we did run into some problems 19:21:12 <fungi> jhesketh: when you see this, thanks for getting it working! 19:21:16 <clarkb> swiftclient is not following semver 19:21:22 <clarkb> so we had to downgrade our swiftclient install 19:21:37 <fungi> ahh, so we hit a backward incompatibility? 19:21:39 <clarkb> because newer version introduced a nonbackward compatible type change 19:21:42 <clarkb> fungi: ya 19:21:58 <fungi> that's unfortunate 19:22:07 <clarkb> we have pinned swiftcleint in os-loganalyze and it should all be fine until we are ready to support the new thing and upgrade though 19:22:12 <clarkb> but that was a really fun evening 19:22:18 <fungi> i can imagine 19:22:21 <anteaya> I remember it 19:22:24 <anteaya> late you were 19:22:36 <fungi> okay, moving on then... 19:22:41 <fungi> #topic Specs approval 19:22:56 <fungi> #info approved "Host a CI systems monitoring dashboard" spec 19:23:06 <fungi> #link http://specs.openstack.org/openstack-infra/infra-specs/specs/deploy-ci-dashboard.html 19:23:13 <fungi> #info approved "JJB 2.0.0 API Changes" spec 19:23:30 <notmyname> clarkb: that is unfortunate and not intended. I'd love follow-up (outside of your meeting is fine) 19:23:30 <fungi> #link http://specs.openstack.org/openstack-infra/infra-specs/specs/jenkins-job-builder_2.0.0-api-changes.html 19:23:37 <clarkb> notmyname: can do 19:23:44 <fungi> #topic Specs approval: StoryBoard Worklists and Boards (SotK) 19:23:51 <fungi> #link https://review.openstack.org/202989 19:24:01 <SotK> This spec has been hanging around without activity for a month and I'd like to get it approved so that we can merge patches to implement it. 19:24:04 <SotK> The work I've done on implementing the spec so far can be seen at http://185.98.149.23/#!/ if people are interested in seeing it. 19:24:21 <jeblair> ooh green 19:25:25 <fungi> any objections to setting a 48-hour deadline to register concern over that spec? 19:25:56 <jeblair> nope, looks like ttx has seen it which was the only thing i was going to offer 19:26:02 <clarkb> none here 19:26:07 <anteaya> no objections 19:26:13 <fungi> #info voting open on "StoryBoard Worklists and Boards" spec until 2015-10-08 19:00 UTC 19:26:17 <clarkb> I am happy for those interested in storyboard to drive their own direction too fwiw 19:26:28 <fungi> same here, glad to see it surviving! 19:26:28 <Zara> great. please let SotK know if you find any bugs in the implementation; he really loves that... 0:) 19:26:34 <fungi> heh 19:26:34 <SotK> thanks people :) 19:26:54 <fungi> #topic Priority Efforts 19:27:02 <fungi> any urgent updates on blockers for the priority efforts listed in our agenda? if no, we've got summit planning coming up and i want to make sure we leave time to discuss 19:27:27 <fungi> going once... twice... 19:27:47 <fungi> #topic Summit planning 19:27:55 <fungi> #link https://etherpad.openstack.org/p/infra-mitaka-summit-planning 19:28:22 <fungi> i see some great ideas on the list so far 19:28:46 <anteaya> remember to add your name ot the list on the etherpad, top right hand corner 19:29:17 <fungi> it looks like so far there's a lot of fishbowl ideas but not many workroom ideas 19:29:29 <anteaya> we like to talk? 19:30:02 <anteaya> if you leave a question or comment on the etherpad please add your name or nick 19:30:11 <Zara> I put mine down as a workroom but I thought it was better suited to a fishbowl, I just didn't want to compete with people for that space since I thought they had a better claim to it. 19:30:37 <fungi> the other thing worth considering is that our 4 workroom slots are in back-to-back pairs 19:30:38 <anteaya> whoever just added the green question to yolanda's proposal 19:30:42 <clarkb> greghaynes: getting noepool image works might be good fishbowl? 19:30:51 <clarkb> greghaynes: and or writing nodepool tests 19:30:58 <anteaya> Zara: looks like you made a good space choice 19:31:02 <jeblair> i don't think zuulv3 is ready for a workroom yet -- however, we have some things i think need to get done before v3 -- greghaynes nodepool dib work and jhesketh's zuul connector work 19:31:15 <jeblair> if those are outstanding, maybe we could attack them there 19:31:21 <anteaya> jeblair: +1 19:31:21 <jeblair> clarkb: good idea! :) 19:31:21 <fungi> so it's _possible_ we might do an extended workroom session if something benefits from 80 (well 90 if you ifnore the break) minutes instead of the usual 40 19:31:24 <yolanda> how is status of masterless puppet? shall we dedicated some time there? 19:31:39 <clarkb> jeblair: I will add them 19:31:51 <crinkle> yolanda: ++ nibalizer ^ 19:31:55 <Zara> anteaya: thanks :) 19:32:00 <greghaynes> Sure, I think the image builder stuff should be done before then but who knows what shows up on reviews 19:32:02 <jeblair> yolanda: yeah, good idea 19:32:21 <clarkb> greghaynes: well maybe we use that time to deploy it across the clouds 19:32:24 <clarkb> greghaynes: and debug when it breaks :) 19:32:35 <greghaynes> Yea, that actually sounds like a good plan A/B 19:32:48 <greghaynes> depending on what state its in 19:33:19 <crinkle> I think infra cloud might be a good workroom session or it could wait till the sprint day 19:33:53 <fungi> as usual, keep in mind that workroom sessions are ideal for mini-sprint type tasks where we identify something that a room full of people can potentially churn out (at least get a great start on if not complete) in 40 (or 90) minutes 19:34:05 <jeblair> crinkle: yeah, depends on if there's something we can accomplish in 40-80 mins with infra cloud... i think of it as more of a half-day thing. 19:34:24 <crinkle> jeblair: i think so too 19:34:26 <anteaya> crinkle: you want to put it on the etherpad somewhere? 19:34:36 <jeblair> crinkle: but if there's a mini-task or plan that needs to be sketched out... that'd work. 19:34:36 <crinkle> anteaya: sure 19:34:41 <fungi> for those who joined us on the irc meetings tracker session, that i think was a mostly perfect example of a good workroom idea 19:35:02 <jeblair> fungi: yeah, we got that 90% done and finished it easily the next week. 19:35:04 <anteaya> crinkle: then it isn't lost, thanks 19:35:31 <fungi> i think it's good to view workroom sessions as "team building" exercises 19:36:17 <fungi> they might be things we could collaborate remotely on, but it's fun and helps with socialization to hack together on a common goal and share that sense of accomplishment in the same room 19:36:30 <jeblair> fungi: and establishing a shared understanding 19:37:50 <fungi> for fishbowl sessions, these tend to be best used for topics which need discussion to arrive at a common consensus or level-set some understanding within the group or cross-project with other stakeholders outside infra, particularly where it might otherwise require a protracted many-weeks-long thread on the ml otherwise 19:38:43 <anteaya> zaro: do you want to keep Gerrit development/upgrade as fishbowl or move to workroom? 19:39:10 <zaro> anteaya: IMO better as fishbowl. 19:39:15 <anteaya> very good 19:39:16 <fungi> we won't necessarily be making final decisions in a fishbowl, but we'll be exchanging positions and getting what we need to be able to have the final follow-up decision discussions on the ml (so that those who can't make it to the summit can still have some input) 19:42:30 <fungi> an important additional detail on the friday sprint time. before lunch we're co-sprinting with the ironic team 19:42:50 <anteaya> and after lunch with qa and release management 19:43:05 <fungi> morning would be a great time for infra cloud since we can potentially collaborate on bifrost/bare metal automation needs 19:43:10 <pleia2> ++ 19:43:26 <clarkb> fungi: and possibly onmetal if that is still something magnum wants to try 19:43:35 <clarkb> though never heard back from them so maybe not 19:43:40 <fungi> and afternoon, yes our usual group therapy with qa and rm 19:43:49 <anteaya> can't miss that 19:43:52 <pleia2> hehe 19:44:02 <jeblair> i never thought about release management being acronymed 'rm'. that's fantastic. 19:44:10 <anteaya> ha ha ha 19:44:11 <jeblair> er, initialed, i guess. 19:44:15 <mordred> fungi: I think it would be useful to be explicit on intent with workrooms ... in Vancouver we had different people with mismatched thoughts, both valid things, but mostly incompatible 19:44:28 <anteaya> that was me 19:44:30 <anteaya> my fault 19:44:45 <mordred> not fault. I just think clarity is good for all 19:44:54 <anteaya> agree clarity is good 19:45:02 <anteaya> but I recall that one and will own it 19:45:12 <fungi> mordred: great point. yes workrooms are where we come prepared to collaborate on some predetermined task, not so much come to learn about something or get presented to 19:45:50 <fungi> so it helps to familiarize yourself with what we're going to work on in advance and be ready to volunteer for subtasks 19:45:50 <pleia2> I think we may also want to consider that we can also split out if we don't feel we need to be there, I kind of thought I *should* but made much better use of my time by splitting off into another smaller group to do other things 19:46:05 <mordred> ++ 19:47:00 <fungi> yep, don't just show up to a workroom session because "it's infra." there will be other places to have conversations or collaborate on unrelated work, and there will also be sessions for other teams going on so it can be a great excuse to cross-pollinate 19:48:28 <fungi> anyway, the goal with the planning topic for today's meeting was to prompt one another for additional ideas and get them onto the etherpad. we'll have one more meeting where we can settle on which topics from the list we're going to schedule into what slots 19:48:50 <anteaya> I think the goal for today was successful 19:49:04 <fungi> so if you think of things between now and next tuesday please do add them to https://etherpad.openstack.org/p/infra-mitaka-summit-planning 19:49:13 <zaro> anteaya: i've changed my mind on that 'Gerrit development/upgrade' would be better as workroom, 'scaling repo creation' better as fishbowl. 19:49:40 <anteaya> zaro: very good, would you like to make that change? 19:49:57 <anteaya> and thanks for adding yourself to co-chair scaling, didn't want to speak for you 19:50:13 <fungi> zaro: ooh, like collaborate on gerrit/jgit hacking and testing during a work session? 19:51:21 <zaro> fungi: I was thinking get people started on gerrit/jgit hacking but if someone already knows gerrit then we can collaborate as well 19:51:49 <anteaya> zaro: you have a few respondants to your request for help 19:51:55 <fungi> zaro: it might be a good idea to find out how many of the people who volunteered to help try to fix those will be summiting 19:51:59 <anteaya> and at least 2 said they would be at summit 19:52:05 <fungi> oh, excellent 19:52:07 <anteaya> persia I remember 19:52:15 <anteaya> and one other, I will check the logs post meeting 19:52:20 * fungi notes that it's hard to forget persia 19:52:26 <pleia2> hehe 19:52:32 <anteaya> so memorable 19:52:47 <zaro> iirc, i think signed8bit is only one that is summiting 19:52:54 <anteaya> and persia 19:53:05 <nibalizer> yes masterless puppet should get some attention 19:53:06 <anteaya> and thanks zaro, yes that is the other person 19:53:36 <anteaya> pleia2: found signed8bit 19:53:38 <pleia2> oh yes, how could I forget 19:54:16 <zaro> anteaya: np. 19:54:22 <anteaya> :) 19:54:31 <fungi> okay, i'll transition this to open discussion for the last 5 minutes since mmmpork had some additional questions about next week's maintenance, but feel free to keep plugging summit session ideas too 19:54:32 <zaro> anteaya: those sessions are already in the right spot. 19:54:43 <anteaya> zaro: wonderful 19:54:44 <fungi> #topic Open discussion 19:55:24 <mmmpork> we might want to have a meeting or discussion about the project rename process 19:55:28 <pleia2> I'm speaking at Grace Hopper next week on our CI system, this is a pretty huge deal and I'm going to be very busy at the conference from Tuesday on http://schedule.gracehopper.org/session/open-source-presentations-3/ 19:55:49 <pleia2> from there I'm off to Japan to do some traveling pre-summit, so see you there? :) 19:55:50 <AJaeger> great, pleia2 ! 19:55:58 <fungi> pleia2: thanks for the heads up, and good luck! (not that you need it) 19:56:12 <SpamapS> speaking of infra-cloud 19:56:15 <fungi> mmmpork: what about the project rename process should we discuss? 19:56:18 <pleia2> fungi: thanks 19:56:28 <jeblair> pleia2: yay! 19:56:34 <SpamapS> just an update.. crinkle and greghaynes pushed a bit and got a mini-cloud up with some hand-massaging of network configs 19:56:39 <mmmpork> 1) what are the pieces that need to be run? 19:56:42 <mmmpork> 2) how will they be run 19:56:50 <mmmpork> 3) are we going to do a test run first? 19:56:52 <SpamapS> the current challenge is to turn that into actual automation and deploy it into the working nodes in HP1 19:57:03 <mmmpork> 4) how will we verify the results before making actual changes to make sure we didn't miss anything? 19:57:15 <anteaya> pleia2: congratulations 19:57:43 <mmmpork> i'm ok with us having a document, an etherpad, or something as long as it's linked on the project wiki and we know who is doing what and when 19:57:49 <anteaya> SpamapS: crinkle greghaynes congratulations 19:58:17 <greghaynes> (it was mostly crinkle) 19:58:20 <pleia2> mmmpork: so project renames are something we do pretty often with the previous stackforge<>openstack churn we used to have 19:58:24 <fungi> mmmpork: so when you say "project rename process" you mean the upcoming maintenance specifically, not just our general project rename process (which we have documented quite thoroughly already)? 19:58:29 <mmmpork> yes 19:58:43 <SpamapS> There should be some patches heading toward system-config with the automation bits. 19:58:46 <mmmpork> since we're going to be doing a bunch of them all at once and having gerrit downtime to do so, correct? 19:59:09 <pleia2> mmmpork: this process will be much the same as the one we've done many times, which I think is why we're pretty casual about this and have our etherpad+docs process 19:59:25 <pleia2> mmmpork: it always requires gerrit downtime 19:59:29 <mmmpork> alright, if you all are comfortable with that then i won't push for it 19:59:33 <crinkle> SpamapS: yes, topic:infra-cloud has relevant patches 19:59:34 <fungi> mmmpork: right, so the documented process is still going to be followed, but we have some additional scripts with this one to pregenerate what we're going to run/merge so that it can be reviewed more easily and further in advance 19:59:59 <fungi> anyway, we're at time 20:00:02 <mmmpork> fungi: do we have that documented anywhere? what scripts are going to be run, where and by who? 20:00:22 <fungi> anyway, we're at time, don't want to hold up the tc 20:00:30 <fungi> mmmpork: we can discuss further in #openstack-infra 20:00:34 <fungi> #endmeeting