19:01:02 <clarkb> #startmeeting infra
19:01:02 <openstack> Meeting started Tue May 29 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.
19:01:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:03 <corvus> howdy!
19:01:06 <openstack> The meeting name has been set to 'infra'
19:01:17 <clarkb> #link https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Agenda_for_next_meeting
19:01:38 <clarkb> There is an agenda, though today I think we may diverge from ti a bit and talk about summit since not all of us were able to be there
19:02:06 <fungi> thanks!
19:02:13 <dmsimard> \o
19:02:28 <clarkb> #topic Announcements
19:02:36 <dmsimard> clarkb: we covered my topic at the last meeting, you were on PTO
19:02:41 <clarkb> dmsimard: thanks
19:02:43 <dmsimard> I'll take it out
19:02:54 <clarkb> I don't have any announcments. Does anyone else have somethign to announce?
19:03:28 <fungi> it was awesome to get to see so many of you last week
19:03:52 <fungi> that's about all i have in the way of announcements ;)
19:04:12 <clarkb> #topic Priority Efforts
19:04:45 <clarkb> mordred mentioned that he will be working on the config management spec that considers the short and long term tasks necessary to get us modernized
19:04:51 <clarkb> I don't think this has been done yet though
19:05:09 <corvus> i'm very eager for that
19:05:46 <clarkb> Mostly just noting it here as a reminder
19:05:56 <clarkb> hopefully we can dig into that soon though. mordred let us know if we can help at all
19:06:19 <clarkb> #topic General Topics
19:06:21 <fungi> storyboard.openstack.org and storyboard-dev deployments of storyboard-webclient are still broken while we work out how to safely perform web content replacement. we got closer last week but as a result had to stick storyboard.o.o into the emergency disable list and untar the latest content by hand
19:06:26 <clarkb> oh sorry
19:06:28 <clarkb> #undo
19:06:29 <openstack> Removing item from minutes: #topic General Topics
19:06:32 <fungi> np. i type slowly
19:06:58 <fungi> that was all i was going to say there
19:07:08 <fungi> i guess i could link your review
19:07:46 <clarkb> well that change is broken as you poitned out
19:08:00 <fungi> #link https://review.openstack.org/570004 Use archive module to handle webclient source
19:08:08 <fungi> yeah, assistance there is welcome
19:08:29 <clarkb> there is probably some puppet pattern we just aren't aware of
19:09:22 <clarkb> #topic General Topics
19:09:45 <clarkb> ianw before I jump into summit things did you want to talk about afs?
19:10:01 <clarkb> your emails were great at keeping everyone up to date I think and thank you for addressing that problem
19:10:28 <ianw> it should be back to ok now
19:10:53 <ianw> it could have been done quicker if i didn't let the kerberos ticket timeout
19:10:57 <fungi> i followed along in #openstack-infra and it seemed really involved. sorry you got stuck with that
19:11:14 <corvus> ianw: did you end up doing the unlock thing auristor suggested?
19:12:03 <ianw> corvus: no ... it was more involved than that, that was just the first step apparently.  it happened late on my friday night so, yeah, was easier to restart
19:12:26 <ianw> auristor was very helpful in finding the thing to delete to uncorrupt things enough to start again.
19:12:34 <corvus> ok
19:12:58 <ianw> i think we would have ended up having to probably delete /vicepa and start the whole thing again without expert overview
19:14:00 <ianw> if we'd like to keep talking about it, pypi is getting ever closer to 2tb which is going to start causing us problems.  but we can also ignore it for a little more too, i upped the quota to 1.9tb
19:14:49 <clarkb> re pypi specifically I think we want to rescan the logs for projects that update frequently and have big packages, update our blacklist then regenerate pypi mirror from scratch
19:15:11 <clarkb> I didn't want to regenerate from scratch until our blacklist of packages was a bit more robust
19:15:24 <clarkb> but I think we save almost a terabyte by regenerating from the current list
19:16:11 <ianw> oh, and an interesting link on debugging volume failures for the future
19:16:11 <corvus> can we just rm the packages from under it?
19:16:18 <ianw> #link https://lists.openafs.org/pipermail/openafs-devel/2018-May/020499.html
19:16:26 <clarkb> corvus: that is another option though we would have to update the indexes too
19:16:31 <clarkb> or we can let them 404 I guess
19:16:43 <corvus> oh, hrm.  maybe better to regen :)
19:17:20 <ianw> they'd still 404 if someone requested it anyway after blacklisting?
19:17:36 <ianw> i.e. it wouldn't be in the mirror?
19:17:50 <clarkb> ianw: yes but the mirror wouldn't claim to have them at all
19:18:00 <clarkb> probably somewhat confusing to users of pip in either case
19:18:52 <ianw> clarkb: do you have a usage script to find things to target?  if not, i can take an action item to get something together
19:19:09 <fungi> regenerating from scratch is cleaner for sure, though perhaps the end result isn't much different functionally
19:19:20 <clarkb> ianw: I don't I mostly just grepped the log and piped through sort for frequency counts
19:19:35 <clarkb> then from there noticed anything machine learning ended up having huge package sizes due to cuda
19:19:42 <fungi> different from scripting up something to delete the existing files for blacklisted projects, i mean
19:21:03 <ianw> ok, i can see about something we can run maybe periodically to give us a view of what's using space
19:21:16 <clarkb> sounds good
19:21:39 <clarkb> Anything else before we talk about the summit?
19:22:33 <anteaya> I'm grateful if anyone feels like reviewing the limesurvey patch: https://review.openstack.org/#/c/557979/
19:22:51 <clarkb> #link https://review.openstack.org/#/c/557979/ limesurvey change
19:22:53 <anteaya> it is almost ready to be a thing
19:23:04 <anteaya> thanks clarkb
19:23:15 <clarkb> #topic Vancouver Summit v2
19:23:51 <anteaya> the weather was great
19:23:52 <anteaya> again
19:24:06 <clarkb> This summit was particularly busy for me with a lot of stuff happening. That said I found it to be very productive which was nice. I'm still sorting out the writing of a summary which I will send to the infra list but figured we could talk about things people had questions about here as well
19:24:25 <clarkb> One major thing that affects this group in particular is Zuul got its day in the spotlight
19:24:34 <clarkb> don't be surprised if you see new users or people curious about it asking questions
19:24:55 <clarkb> Also there are stickers, corvus has a collection of them
19:25:23 <fungi> heh, yeah i did _too_ good of a job booking up every last minute of my schedule for last week. at one point i had to sprint from a forum discussion to the lightning talk stage and then sprint back again as soon as the talk concluded
19:25:31 <fungi> clarkb can verify i was out of breath
19:25:42 <fungi> thankfully diablo_rojo_phon was there to pick up my slack
19:25:47 <clarkb> Other areas of interest include openstack now has a concrete proposal to deprecate python2, expect email to the dev list from dhellmann to finalize that plan
19:26:06 <clarkb> OpenDev resulted in some great interaction between communities and organizations around doing CI/CD
19:26:37 <clarkb> I heard the zuul workshop in particular was really well done. Thank you to pabelanger for that (I think he had help as well but unsure who was helping)
19:26:48 <corvus> mnaser helped pabelanger
19:27:29 <fungi> was it well-attended? sorry i wasn't in that one (forget what conflict i had during that slot)
19:27:37 <corvus> sold out
19:27:38 <clarkb> A few of us were able to meet with Kata and make a plan to run some of their jobs with zuul with the hope for a moving to zuul long term
19:28:49 <clarkb> The first contact group has also been working to better document what companies need to allow their employees to do in order to work upstream successfully. Things like send email to the mailing list
19:29:21 <clarkb> On the subject of mailing lists there seems to be a strong push to consolidating our mailing lists so don't be surprised if openstack drops down to a much smaller number of active mailing lists in the near future
19:29:55 <clarkb> But I don't want to summarize everything in this meeting. Will instead ttry to write an email that covers the broad topics. Instead are there specific topics people want to go over whether you were at the summit or not?
19:30:07 <fungi> yeah, i'll have something out to the ml soon as a summary for that first contact sig session
19:31:26 <corvus> https://review.openstack.org/570087 is an outcome of an opendev session -- hopefully we'll hook up our zuul to opendaylight's gerrit so we can do cross-testing
19:31:47 <corvus> i set up accounts and updated our infra-root password store; waiting on odl permission changes now
19:31:58 <fungi> that'll be awesome
19:32:14 <clarkb> #link https://review.openstack.org/570087 enable cross testing of changes in OpenDaylight's gerrit with OpenStack's gerrit via Zuul
19:34:18 <fungi> one thing we informally made a little progress on at the forum, we're trying to work out a process for the discussions we need to have around renaming the infra team effort. like most renaming discussions it's complicated by legal review and finding ways to avoid being put at a disadvantage during domain name acquisition
19:36:04 <clarkb> Overall there seemed to be a willingness to adapt the way we and others do things to be more flexible as our users' needs change. The mailing list updates, talking to other gerrits, possible renaming of infra to be a bit more neutral to non openstackers, and so on. I don't expect everything to change all at once and this should happen with community input.
19:36:32 <clarkb> If these types of changes concern you or you otherwise have input definitely make sure your voice is heard. I'm happy to listen to feedback if people are comfortable putting it out in public as well
19:37:45 <corvus> i'm happy with the approach of sharing names privately when necessary, and otherwise discussing the process in public
19:39:46 <clarkb> on that specific topic, I was asked to send email to the foundation that can be used to engage the legal team re the specific name fungi and I had in mind. The idea being if there aren't legal concerns we can circulate it among the team and see if we have consensus and move forward from there. The concern being the cost involved with the name and that potentially changing.
19:40:38 <fungi> yeah, i hate the cloak-and-dagger bs around naming things
19:41:18 <corvus> it's too bad the name-everything-a-uuid movement never caught on
19:41:24 <clarkb> indeed
19:41:33 <anteaya> ha ha ha
19:41:36 * fungi registers all possible uuids
19:44:01 <clarkb> Another major item worth mentioning was that opendev attendees seemed comfortable using openci.io to continue the collaborative discussion around ci going forward
19:44:21 <clarkb> feel free to join the mailing list there and otherwise get involved
19:44:21 <anteaya> awesome
19:44:35 <fungi> yes, i thought those sessions were great for sharing ideas across communities
19:45:33 <fungi> i got several thanks from random people for the ci/cd security fundamentals discussion (even though i basically just got voluntold to moderate it due to lack of actual volunteers)
19:46:11 <fungi> proof that just getting people in a room labelled with a broad topic and asking them to riff on that is helpful
19:46:25 <anteaya> thanks for the ci/cd security fundamentals discussion fungi
19:46:33 <fungi> heh, you're welcome!
19:47:06 <clarkb> anything else before I open the floor to open discussion?
19:48:25 <clarkb> #topic Open Discussion
19:49:22 <anteaya> and I'll agree with fungi's statement earlier, it was very nice to see so many of you again
19:50:50 <clarkb> I'll admit it is nice to not be dealing with jet lag :)
19:51:30 <anteaya> clarkb, :)
19:51:34 <fungi> i'm dreading watching the video of the tech elections lightning talk and seeing that i was probably doing an even worse delivery than i thought at the time
19:51:44 <clarkb> Doesn't sound like we have much else to cover today. I'll end the meeting in a minute or two if nothing else comes up.
19:51:49 <anteaya> I just got home last night, drove around the island for a bit, so I'm lagging today
19:52:11 <anteaya> fungi, I'm sure you were fine
19:52:24 <anteaya> your elecution is always very good
19:52:57 <anteaya> clarkb, I have nothing else, just chatting
19:53:05 <corvus> fungi's electrifying election elocution
19:53:10 <anteaya> ha ha ha
19:53:18 <clarkb> yesterday was a US holiday and we just had the summit so I expect this to be a very quiet week
19:53:21 <fungi> heh
19:53:40 <fungi> yeah, i expect i'll be lagging all week. travel takes a lot out of me, but a week solid of being in sessions even moreso
19:54:07 <clarkb> Thanks everyone
19:54:09 <clarkb> #endmeeting