19:01:21 #startmeeting infra 19:01:22 Meeting started Tue Jun 5 19:01:21 2018 UTC and is due to finish in 60 minutes. The chair is clarkb. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:23 o/ 19:01:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:01:26 The meeting name has been set to 'infra' 19:01:34 #link https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Agenda_for_next_meeting 19:01:42 Hello 19:01:45 #topic Announcements 19:02:45 I don't really have any announcments 19:03:24 #topic Actions from last meeting 19:03:32 #link http://eavesdrop.openstack.org/meetings/infra/2018/infra.2018-05-29-19.01.txt minutes from last meeting 19:04:43 Don't see any explicit actions on the last meeting notes, but the change to talk to opendaylights gerrit is still up. https://review.openstack.org/#/c/570087/ 19:05:37 #link https://review.openstack.org/#/c/572451/ Is another item that we talked about last week. Basically start talking to kata projects with zuul 19:05:39 were we waiting for something before approving that? 19:05:47 570087 i mean 19:05:50 #link https://review.openstack.org/#/c/570087/ Talk to opendaylight's Gerrit with our zuul 19:05:57 fungi: we needed them to give our account stream access perms 19:06:04 oh, right 19:06:05 I believe zxiiro took care of that 19:06:14 thanks zxiiro! 19:06:21 np :) 19:06:31 so i can just approve it then 19:06:35 I think so 19:07:11 #topic Specs approval 19:07:42 I haven't really had time to catch up on the state of specs post summit. I don't think mordred has pushed up the new config mgmt future spec yet but promises me the draft was oepn in vi locally :) 19:07:56 are there any specs we should be looking at? 19:08:29 i can't see mordred's screen from here, unfortunately 19:08:40 I love it when updates include what files mordred has open on his laptop 19:09:14 #topic Priority Efforts 19:09:22 Let's start with storyboard. 19:09:28 #topic Storyboard updates 19:09:38 openstack-helm and magnum are migrating on friday 19:09:43 I think we got storyboard web client installing properly on the web server finally 19:09:48 yeah, that too 19:10:00 it's updating normally now from what we can tell 19:10:13 Looks like Keystone has kicked things off on the dev list discussing things since I finally got around to test migrating them to the dev site 19:10:38 and today's tripleo weekly newsletter indicates they're evaluating it 19:11:05 TC is looking at using it for tracking community goals as well? or did I misread that discussion 19:11:05 and something about the ui squad looking at tracking specs/epics in there 19:11:17 no, the tc has been doing that for a cycle already 19:11:25 ah gotcha 19:11:36 there is recent talk about migrating stuff to it though? just older items maybe? 19:11:38 they're talking about whether to try to move the previously gerrit-tracked lists of goal completion artifacts into storyboard for historical purposes 19:11:43 roger 19:11:48 clarkb, yeah older items I think 19:11:56 but current goal completion is already being tracked in sb 19:14:00 i think that's all for sb updates this week? 19:14:29 how is outreachy going? 19:14:34 has that started yet? 19:14:50 clarkb, she is just getting started- finished finals a week or so ago. 19:15:00 i saw our intern pop into #storyboard briefly last week 19:15:10 * diablo_rojo has a todo to reach out and see if she needs anything since last week 19:15:34 may 30, had some tox questions it seems 19:15:37 cool, thanks for the update. Sounds like good progress and hopefully with web client being deployed reliably we'll start seeing those new patches soon :) 19:15:56 #topic Future Config Management 19:16:16 as mentioned earlier I don't think the latest iteration of the spec for this is up yet, but Monty promises it is in progress :) 19:16:33 any thoughts/cocnerns related to this particular after decompressing from summit? 19:17:06 i've seen a lot of patches from cmurphy for puppet 4 uplift 19:17:19 ya i'm starting to run out of things to do 19:17:26 reviews welcome 19:17:31 though i haven't had a chance to review them, so not sure how that's going 19:17:39 #info cmurphy's puppet 4 uplift changes need reviews 19:17:44 sounds like nobody else has either. sorry! :/ 19:17:53 I was trying before summit things happened 19:18:13 its mostly a matter of getting testing dnoe and updating manifests right? 19:18:22 i've found a lot of the modules don't have any beaker tests at all, not sure how much time i want to spend on writing them :/ 19:18:40 if the updates are trivial (which many I've seen have been) we probably don't need to add tests 19:19:28 I'll do my best to dig into that soon (hopefully today) 19:19:50 we can migrate them one at a time so i feel like we can do the -dev servers first as canaries instead of writing complete tests 19:20:00 that seems reasonable 19:20:03 i can take a look too ... i've found writing the rspec tests quite helpful in bringing up xenial support in a few cases, for sorting out initial issues 19:20:06 that sounds like a swell plan 19:20:46 ianw: ++ i was looking at adding tests for puppet-askbot and found it does not work on xenial yet :/ 19:21:16 #info writing rspec tests has been useful for finding out if puppet manifests run on Xenial 19:22:02 cmurphy: i have a review set out for that https://review.openstack.org/#/c/558991/ & down 19:22:38 ianw: :D 19:22:43 i'll take a look 19:23:18 thanks, upgrading that has been on my todo, there's some discussion on it on openstack-dev somewhere 19:23:36 is wiki.o.o properly puppetized yet? that one might end up being tricky if not 19:23:42 cmurphy: I don't think so :/ 19:23:49 :'( 19:24:31 #link http://lists.openstack.org/pipermail/openstack-dev/2018-April/129078.html 19:24:31 it's not, no 19:24:36 ^ re the askbot 19:24:44 wiki-dev.o.o is using the puppet-mediawiki module we have, at least 19:25:09 however it's not fully functional yet (some odd interaction between the openid plugin and some other plugin, maybe mobile view?) 19:25:37 wiki.o.o is a cname to wiki-upgrade-test.o.o (don't ask, long story) which is not puppeted 19:25:49 some easy reviews here https://review.openstack.org/571022 will let us close out http://specs.openstack.org/openstack-infra/infra-specs/specs/puppet_4_prelim_testing.html i think 19:26:34 so anyway, i guess on a positive note you can't break wiki.o.o by updating puppet-mediawiki at least 19:26:38 #link https://review.openstack.org/571022 will close out http://specs.openstack.org/openstack-infra/infra-specs/specs/puppet_4_prelim_testing.html 19:27:12 alright anything else before we move on? 19:27:36 fungi: i guess that's probably true 19:27:50 o/ 19:28:03 mordred: last chance re config management before we move on to othertopics 19:28:13 (sorry, in flight, a little late getting to altitude) 19:28:31 no - move on - I've got it open in my editor but still not finished. sorry for the delay 19:28:54 mordred: no problem at least I too feel generally delayed the last week and a half 19:29:00 #topic General Topics 19:29:09 pabelanger: you around? 19:29:25 yah 19:29:53 pabelanger: was going to mention things like Gerrit 2.15 upgrade and transitioning test resources to bionic 19:30:00 pabelanger: do you want to fill us in there or should I? 19:30:15 sure 19:31:22 I'm sure most people have heard, but over the next few months, I expect my level of upstream work to be less. So, I can't sign up to do the gerrit / bionic tasks right now. I'd figure we should see if anybody else can do it 19:32:09 for bionic, that is migrating PTI jobs from xenial to bionic. Which should be not that bad 19:32:17 thanks for all your hard work over the years, and glad it doesn't sound like you're leaving us completely at least 19:32:28 and for gerrit, I think coming up with a plan with clarkb and fungi to decide if we do database migration now or later 19:32:46 fungi: yah, not leaving, just tasked with downstream things now 19:32:48 The major tasks I identified we'd probably need new volunteers for are the Gerrit upgrade, bionic transition of testing, general control plane upgrade push, and moving our zuul and nodepool install to the new zk cluster. I don't think we need volunteers for all that right now but if you are interested let me or other infra roots know 19:33:24 pabelanger: ya we need someone to research that and maybe even test it. I tried to do some research and it seems like we might be able to do the db migration incrementally 19:33:43 btu that needs testing and verification (but if that is possible we may have a less painful upgrade path) 19:33:55 I'll be able to do some work, just not drive it sadly 19:35:06 i'm hoping some of our regulars who aren't around for the meeting might decide they're interested in one (or both) of those 19:35:26 today's is not an especially crowded meeting 19:35:31 #info Gerrit upgrade, Bionic transition for test platform, control plane upgrade, and Zuul/Nodepool zk cluster migration need new volunteer(s) to drive the work 19:36:00 pabelanger: and yes thank you for all the hard work. Very much hope you'll be back at it soon 19:36:18 ++ 19:36:54 if anyone is interested in these tasks feel free to reach out to myself 19:37:16 #link https://review.openstack.org/#/c/571520/ update job config requirements for PTI jobs to allow them to live in project repos 19:37:38 This is an item that came about after being pinged quite a bit for trivial project-config changes that were blocking various projects that could otherwise be in repo 19:37:55 ++ 19:38:03 ajaeger wasn't able to join us during the meeting but asked that we get group consensus on it before moving forward 19:38:22 to me this sort of feature is one of the major reasons zuulv3 happened. I think we shoudl take advantage of that feature 19:38:28 we need feedback from release, requirements and tc ideally 19:38:45 fungi: yup those were groups ajaeger brought up. If the infra team isn't opposed I can reach out to those groups 19:39:03 i'm strongly in favor of the infra team ceasing to be police of the openstack pti, but having buy-in would be great 19:39:06 * mordred thinks it's a great idea 19:39:36 I've gone ahead and WIP'd it until we have external buy in. If you are in favor can you leave a positive vote on the change? 19:40:12 especially with us moving to a separate organization under the osf, expecting us to be intimately familiar with job policies for all the projects we host seems wrong 19:40:15 that way I can show those groups that infra itself is onboard pretty easily 19:41:20 ok, anteaya I think you are up with the limesurvey update 19:41:38 okay thanks 19:41:52 so its a thing https://survey.openstack.org/ 19:42:08 so if you go to https://survey.openstack.org/admin you should be able to log in via openstackid 19:42:14 and create your own surveys 19:42:21 are all surveys public? 19:42:27 good question 19:42:29 at the moment my openid is the only superadmin but i can set anyone else's accounts up with full privileges to manage it through the webui 19:42:40 my understand is that no, they are not public 19:42:58 but that may be something I need to look more closely at in the settintgs 19:43:06 and yes, this should still be considered a sort of proof-of-concept until we try to use it for a few things and have a better grasp of what it can and can't do 19:43:07 ok I think its fine to have limtied surveys and open surveys. I was just curious if this had an opinion on that 19:43:08 currently one needs to recieve a token to take a survey 19:43:12 fungi: gotcha 19:43:19 "beta" testing phase if you will 19:43:35 now tokens can be sent to emails 19:43:52 or reused, but if reused I think association with the submitter is lost 19:43:59 and anyone can log into the /admin interface with their openstackid and start creating surveys too 19:44:07 I'm gathering feedback on how this all works 19:44:18 \o/ 19:44:19 and will be poking about myself now that it is alive 19:44:29 a few known gotchas... 19:45:02 biggest is that if you authenticate it autocreates an account for you (good!) but currently hard-codes the initial name and e-mail address in your profile to dummy values 19:45:35 we could stand to have someone write a very simply php hook to pass those values dynamically from the apache envvars provided by mod_auth_openid 19:45:43 s/simply/simple/ 19:45:57 at least the documentation for limesurvey makes it look fairly extensible in that regard 19:46:10 I would offer to do that - but I think that would not e positive for anyone 19:46:27 also yeah, it still needs some theming and whatnot (it's basically all the default ui options for the moment) 19:46:30 mordred: because it wouldn't be completed you mean? 19:46:34 fungi: any chance we could trick jimmy or friends into writing a hook? 19:46:39 anteaya: well, and also I'm terrible at php 19:46:45 mordred, ah 19:46:46 mordred: that was my first thought as well 19:47:23 at least we have a puppet class and instructions for setting up a limesurvey instance from scratch, so testing additions to it should be trivial for anyone with a vm to burn 19:47:26 aslo docs patch for limesurvey admin is up 19:47:35 #link limesurvey admin docs https://review.openstack.org/#/c/571536/ 19:47:56 I need to respin to take in fungi and ianw's comments 19:48:02 thanks for providing them 19:48:48 mine was just minor and only if we need a respin, but i see fungi has added more substantive comments 19:49:09 ianw glad to have your review, thanks for taking the time to look at it :) 19:49:41 my comments were really more the result me testing some theories about how to streamline the initial setup 19:50:06 that's all I had on this 19:50:09 thank you 19:50:16 and tahnk you for helping get that running 19:50:24 yes, thanks anteaya! 19:50:27 thanks for letting me 19:50:32 and thanks to all who reviewed the patches 19:50:36 and thanks for all the heavy lifting fungi 19:50:55 once we get the documentation change merged we can likely push a change to move the spec to completed 19:51:02 fungi: sounds like a plan 19:51:06 awesome 19:51:21 #topic Open Discussion 19:51:26 though i expect the document will grow some basic pointers for people creating/running surveys 19:51:43 fungi, the limesurvey docs cover that really well 19:51:44 that should be fine, the service is up and useable and bulk of work is done 19:51:58 before our hour is up anything else? 19:52:21 ummm, I've been planting my garden? 19:52:43 since it is a hot topic over the last couple days, I will mention that I don't think the infra team needs to do anything re github being bought. It is still a proprietary service run by a company that hasn't changed 19:53:06 ++ 19:53:26 I'll happily point people to alternatives if they are interested in moving but we just mirror there 19:53:49 (and that is something we've long grumbled about, but this doesn't change the factors of that grumbling imo) 19:53:55 there was an interesting announcement posted to the debian-devel-announce ml about creating a "data protection team" to act as a point of contact over gdpr questions/queries 19:54:05 i'd link it here but the web archive seems to be lagging 19:54:25 ahh, there 19:54:50 #link https://lists.debian.org/debian-devel-announce/2018/06/msg00000.html Data Protection team delegation 19:55:19 topics to keep in mind, since i know we've sort of rhetorically asked ourselves what the gdpr means for the services we're running 19:55:44 fungi: ++ 19:56:23 in general gerrit seems to be of the opinion that the data in git fits under the it can't be changed exemption and our service logs are rotated pretty aggressively in general. Which probably leaves accounts in $service specifically 19:57:43 using openids helps a lot there too 19:58:12 since we're not maintaining databases of passwords at the moment 19:58:23 there has been confusion about what you can log and my reading of the rules is you can log for a reasonable period of time to debug issues in services 19:58:32 it does - ut we copy info from the openid into databases for some of the services (like gerrit) - might not be a bad idea to figure out how to remove such info if asked - or whether anything we copy in is covered 19:58:32 so it looks like one of the requirements is appointing a point of contact for inquiries, yes? 19:58:36 so unlike some of the patches pushed to openstack you don't have to remove all the info entirely 19:58:39 you just have to rotate it out in logs 19:59:32 mordred: fungi there is also the potential argument (and I'm not a lawyer) that our four opens gives us a more global exemption 19:59:46 also - as anteaya said - just a point of contact for people to ask for thinngs - even if the response is "that log will be rotated kthxbai" 19:59:50 since the history or change is so important and that happens out side of git 20:00:10 anteaya: mordred I think if you are a company you must do this 20:00:20 looks like that is it 20:00:24 thanks clarkb 20:00:26 its a bit more nebulous in the context of community run infrastrucutre but a reasonable thing to have either way 20:00:35 maybe if anyone complains, we'll just forward them to jbryce :) 20:00:51 and we are at time 20:00:53 thank you everyone 20:00:58 #endmeeting