19:01:02 <clarkb> #startmeeting infra
19:01:03 <openstack> Meeting started Tue Aug 28 19:01:02 2018 UTC and is due to finish in 60 minutes.  The chair is clarkb. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:07 <openstack> The meeting name has been set to 'infra'
19:01:12 <clarkb> #link https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Agenda_for_next_meeting
19:01:22 <clarkb> #topic Announcements
19:01:40 <clarkb> The OpenStack Rocky release happens on Thursday
19:01:44 <clarkb> at least this is what I have been told
19:01:58 <Shrews> o/
19:02:19 <clarkb> fungi should have a new gpg key for that pushed out shortly so be on the lookout for that and be ready to sign it if you can
19:02:33 <fungi> yes, it's generated
19:02:51 <fungi> 0xD943D5D270273ADA is the master key id
19:03:25 <fungi> haven't pushed it up to the keyserver network yet nor added the symmetric decryption passphrase for it to the secrets list
19:04:00 <fungi> there are some nuanced differences in the behavior, configuration and options for the gnupg version in bionic, so with the move to bridge.o.o i have some notes i'll work into a process documentation update
19:04:18 <clarkb> that seems like a fairly common activity around this transition
19:04:23 <fungi> yup
19:04:33 <fungi> also worth noting, the validity period is set to 10 months
19:04:43 <clarkb> because we expect the next cycle to be longer?
19:04:44 <fungi> because the stein cycle is a tall one
19:04:47 <fungi> er, long one
19:04:50 <fungi> long pour?
19:04:53 <fungi> anyway, yes
19:05:15 <fungi> release is scheduled for april
19:05:48 <clarkb> #topic Actions from last meeting
19:06:01 <clarkb> #link http://eavesdrop.openstack.org/meetings/infra/2018/infra.2018-08-21-19.02.txt Minutes from last meeting
19:06:11 <clarkb> fungi: thank you for running that last week
19:06:21 <fungi> np
19:06:26 <clarkb> I don't see any explicit actions but appears that we did a lot of talking around testing of the ansible work
19:06:45 <clarkb> I've been reviewing changes that add said testing today so thank you for getting that in
19:07:28 <clarkb> which is probably a good lead into our config management update priority effort
19:07:33 <clarkb> #topic Priority Efforts
19:07:46 <clarkb> #topic Update Config Management
19:08:12 <clarkb> I'm still sort of catching up on the changes that have happened around this from last week. I've jumped onto getting launch node to work again though
19:08:35 <corvus> clarkb: i think your email basically laid out our next work items :)
19:08:56 <corvus> clouds.yaml, cloud launcher,  etc
19:08:56 <fungi> it was nice and detailed
19:09:08 <clarkb> #link https://review.openstack.org/#/c/596873/ and its parent will get launch node minimally working again
19:09:18 <clarkb> that change has reviews but its parent needs them too
19:10:00 <clarkb> #link http://lists.openstack.org/pipermail/openstack-infra/2018-August/006108.html Thread on getting launch node working
19:10:11 <corvus> change https://review.openstack.org/596849 from Shrews is the last base role we identified last week.  i just +3d it
19:10:15 <clarkb> the major outstanding item now is getting puppet installed so that puppet can run on new nodes that are still puppeted
19:10:37 <clarkb> ianw: ^ has been working on that and I think after a rebase and squash with some minor test changes we should be ready to get that in?
19:11:27 <ianw> one comment from corvus on the integration roles but we can follow up outside meeting i think
19:11:54 <clarkb> #link https://review.openstack.org/#/c/596913/ is another important change that will allow ansible to work on xenial as we transition off trusty
19:12:05 <clarkb> I maybe need ot update the topic on these changes
19:12:11 * clarkb scribbles a note to do that after the meeting
19:12:32 <corvus> oh yeah, using update-cfg-mgmt on these would be helpful
19:12:40 <clarkb> corvus: does that mean we are near to having nothing in server.pp?
19:12:53 <corvus> clarkb: i think so
19:12:58 <clarkb> neat
19:13:20 <Shrews> clarkb: 913 might need changes from 849 for unattended upgrades
19:13:22 <corvus> looks like automatic upgrades and afs client are the last
19:14:05 <clarkb> Shrews: ok I can rebase that if necesary (will look after the meeting)
19:14:35 <corvus> ianw: i'm not sure what's needed to remove afs from server.pp... you've been mostly looking at wheel builders, but i imagine this also affects zuul executors, files.o.o, etc...
19:14:56 <ianw> yes, that was the plan, i can look at the control-plane side
19:15:45 <corvus> maybe the afs client role should be in per-hostgroup playbooks?
19:16:14 <corvus> so instead of moving it from server.pp to base.yaml, we move it to zuul-executor.yaml or whatever...?  it'll slow down the removal of server.pp slightly, but i think makes more sense
19:17:04 <ianw> as in staging the rollout?
19:18:00 <ianw> that's definitely a good idea.  we just had the wheel jobs change approved today, so i will first validate the roles in action there, and then can add them as we progress
19:18:19 <clarkb> once I've got the launch node python and puppet intsallation stuff mostly behind me I'll hopefully have time to help with cloud launcher and clouds.yaml changes
19:18:56 <pabelanger> yah, I'm starting on clouds.yaml change, just getting some testinfra things in place
19:19:08 <pabelanger> should be able to get eyes on reviews for tomorrow
19:20:26 <clarkb> anything else on this topic?
19:20:43 <clarkb> seems like its mostly just working through the todo list (and we are doing that \o/)
19:21:22 <clarkb> #topic Storyboard
19:21:46 <clarkb> fungi: diablo_rojo: I have not caught up on storyboard happenings over the last week. Anything to bring up here?
19:22:06 <diablo_rojo> Two migrations of official projects + one unofficial project this Friday.
19:22:07 <fungi> a few migrations probably going down friday
19:22:20 <fungi> yeah, freezer, searchlight and slogging
19:22:33 <diablo_rojo> Lots of SotK's things got merged. Only a few open reviews which is cool.
19:22:39 <diablo_rojo> Time to get some new work in.
19:22:53 <AJaeger> ;)
19:22:54 <diablo_rojo> I have one, fatema_ has one, SotK has a few.
19:22:57 <fungi> particularly lots of webclient fixes/improvements
19:23:00 <fungi> got merged
19:23:00 <diablo_rojo> AJaeger, :)
19:23:20 <diablo_rojo> That's all I have there.
19:23:38 <fungi> yep, i didn't have anything new to report either
19:23:41 <clarkb> fungi: I assume that we are still deploying those updates successfully?
19:23:45 <clarkb> (so users will see them)
19:23:55 <fungi> correct
19:24:11 <fungi> as of a while back, the webclient about page also mentions the git commit
19:24:41 <fungi> #link https://storyboard.openstack.org/#!/page/about
19:24:48 <fungi> "Webclient version: e6bf69d..."
19:25:08 <fungi> so it's easy to check whether it's current
19:25:14 <clarkb> that is helpful thanks
19:26:09 <clarkb> #topic General Topics
19:26:17 <clarkb> This may end up being a short meeting.
19:26:44 <clarkb> The motivation for getting launch node working is mnaser is giving us access to a new vexxhost region to burn in with nodepool
19:27:00 <fungi> i hear it has some very recent openstack
19:27:00 <clarkb> I'm working on spinning up the in region mirror there (which reuired launch node)
19:27:12 <clarkb> I think mnaser said it is running rocky
19:27:22 <fungi> right on!
19:27:37 <corvus> not stein?
19:27:46 <corvus> so old
19:27:50 <clarkb> mostly an FYI that this is happening and keep an eye out for odd nodepool behavior
19:28:10 <AJaeger> a low prio task - we have a couple of repos that look dead, see http://lists.openstack.org/pipermail/openstack-infra/2018-August/006058.html - anybody interested to review and suggest which ones to retire?
19:28:37 <corvus> oh, flashing back to the ansible topic -- i'm also resuming work on the per-project ssh keys work for zuul, as it may be helpful during our ptg work
19:29:00 <clarkb> #link http://lists.openstack.org/pipermail/openstack-infra/2018-August/006058.html list of openstack-infra repos which have been idle and may need retirement
19:29:08 <clarkb> AJaeger: I can take a look. I think the vinz repos for sure can be retired
19:29:34 * AJaeger is happy to help retiring...
19:29:38 <AJaeger> thanks, clarkb
19:29:39 <corvus> i'd work from your list at the bottom, for starters.
19:30:02 <clarkb> fungi: you had a winterscale topic item, I think this was from last week but probably worth bringing up again today? Basically there is an email thread now where your feedback would be appreciated
19:30:19 <fungi> yeah, that's pretty much the long and short of it
19:30:21 <clarkb> #link http://lists.openstack.org/pipermail/openstack-infra/2018-August/006075.html winterscale naming thread, please provide feedback
19:30:22 <corvus> (git-restack is not dead, it's just perfect)
19:30:28 <fungi> would be nice to pick a name before the ptg
19:30:30 <clarkb> corvus: ha
19:30:32 <clarkb> fungi: ++
19:30:41 <clarkb> I think if we can decide on a name that will make many people happy
19:30:45 <fungi> oh, speaking of idle, how do people feel about knocking out a new git-review release?
19:30:52 <fungi> #link http://lists.openstack.org/pipermail/openstack-infra/2018-August/006084.html git-review tagging schedule
19:30:56 <corvus> [not actually kidding :)]
19:31:04 <clarkb> fungi: I'm happy that gerrit reached out about it
19:31:04 <fungi> i didn't see much from the infra team on that thread
19:31:09 <clarkb> fungi: I htink we should go ahead and do that
19:31:13 <corvus> sounds important
19:31:17 <fungi> yes
19:31:21 <AJaeger> corvus: that's why I didn't list it at the bottom ;)
19:31:39 <fungi> they tracked down an openstack mailing list (granted not the one we watch closely, but still)
19:32:06 <clarkb> will probably make electrofelix and others happy to have some bug fixes included as well
19:32:25 <fungi> yeah, as noted i at least want to get the release notes change in if anyone is up to approve
19:32:26 <clarkb> fungi: maybe do it after the openstack release just to avoid tripping that up with new behavior?
19:32:48 <fungi> sure, official rocky release is two days away anyway
19:32:59 <clarkb> fungi: do you have a link to the change you want in?
19:33:00 <corvus> the lesson here, btw, is *not* to adjust your code until required by deprecation.  you might think the lesson is the opposite -- but we're only in this situation because we adopted refs/publish, because it was supposed to replace refs/for.  :)
19:33:03 <clarkb> I don't see it on the -infra list thread
19:33:12 <clarkb> corvus: indeed
19:33:15 <fungi> but to be honest, i don't think any contributions people push to gerrit tomorrow are likely to be in the initial rocky tag anyway ;)
19:33:37 <ianw> I haven't had too much more feedback on the 3rd-party CI spec.  I will probably re-frame this as a "help wanted" spec, because I don't think anyone one single person is willing to drive it ATM (if you are, speak up).  But i'll be glad if it is in a sanctioned state such that someone with interest can work from it
19:33:48 <fungi> clarkb: it's linked in the last paragraph of the message i linked
19:33:52 <clarkb> ianw: ++
19:33:54 <corvus> ianw: that sounds appropriate to me
19:34:10 <fungi> #link https://review.openstack.org/593670 Start using reno for releasenotes
19:34:13 <clarkb> hrm my email client has broken up the threads looks like
19:34:53 <fungi> oh, though that's actually one change in a series
19:35:18 <fungi> the release notes themselves are in 593671
19:35:30 <fungi> no, ignore me, that one's abandoned
19:35:49 <fungi> it was squashed into 593670
19:36:35 <clarkb> #topic Open Discussion
19:37:05 <fungi> ianw: yeah, i'd like to think that someone interested in running/upgrading to zuul v3 for their third-party ci system might pick up some tasks from that spec
19:37:09 <clarkb> It didn't make it to the official agenda because I'm still catching up after a week off, but a reminder that the PTG is coming up ina few weeks
19:37:55 <clarkb> collecting agenda thoughts at https://etherpad.openstack.org/p/infra-ptg-denver-2018 and if you want to join us for biergarten dinner whilke there please put rough availability on https://ethercalc.openstack.org/fgry72wu42qw
19:38:05 <clarkb> I'll send out a note with a night for that week before PTG probably
19:38:21 <clarkb> in two weeks
19:38:24 <clarkb> wow where did time go
19:39:26 <clarkb> floor is open for any other items, but I'll probably end the meeting at 19:45 UTC if nothing comes up and we can go eat lunch or breakfast 15 minutes early :)
19:39:29 <clarkb> or dinner!
19:39:51 <Shrews> Oh, for those interested, we are due to begin pushing zuul and nodepool container images to dockerhub any moment now (next merge for zuul and after https://review.openstack.org/597171 lands for nodepool)
19:40:38 <pabelanger> https://hub.docker.com/r/zuul/
19:40:42 <pabelanger> looks like some are live?
19:40:45 <Shrews> next we'll add infra edge computing somehow
19:40:54 <corvus> yeah, i enqueued a ref earlier
19:40:56 <Shrews> pabelanger: oh neat!
19:41:10 <pabelanger> I'm totally spending some time now to try them out
19:41:39 <clarkb> oh another PTG item is Friday is a "free" day for infra.  Ithink TC may be doing stuff on Friday but thought it gave us an opportunity to find something fun to do together if we want
19:41:49 <AJaeger> can we add some descriptoins for the images?
19:42:12 <pabelanger> AJaeger: hopefully not via webui
19:42:19 <AJaeger> :)
19:42:24 <corvus> wow that's a lot of screen real estate for almost no content
19:42:34 <clarkb> pbrx can probably grab the description out of the setup.cfg and publish that?
19:43:12 <Shrews> i don't thing setup.cfg provides per-entrypoint descriptions? so that might be a bit difficult
19:43:17 <fungi> note that the "something" the openstack tc members are doing on ptg friday is holding sessions for which anyone is welcome to attend and participate
19:43:43 <clarkb> fungi: is there an agenda for those yet?
19:43:54 <clarkb> (I wasn't sure of the details just saw a room blocked off for TC)
19:43:56 <fungi> #link https://etherpad.openstack.org/p/tc-stein-ptg
19:44:02 <clarkb> thanks!
19:44:36 <corvus> sunday, otoh, is a day of internal reflection for the tc
19:45:12 <fungi> i wanted sunday to be wide open too, but there wasn't much space available on short notice
19:45:40 <fungi> this was as an alternative to trying to get together around some other conference where even fewer of the tc members were likely to be around
19:46:16 <clarkb> alright its 19:45 UTC I'll go ahead and call it a meeting we can follow up in the infra channel and on the infra list. Time for food :) thank you everyone.
19:46:33 <clarkb> #endmeeting