16:00:00 <nirik> #startmeeting FESCO (2017-05-26)
16:00:00 <zodbot> Meeting started Fri May 26 16:00:00 2017 UTC.  The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:00 <zodbot> The meeting name has been set to 'fesco_(2017-05-26)'
16:00:00 <nirik> #meetingname fesco
16:00:00 <zodbot> The meeting name has been set to 'fesco'
16:00:00 <nirik> #chair maxamillion dgilmore jwb nirik jforbes jsmith kalev sgallagh Rathann
16:00:00 <zodbot> Current chairs: Rathann dgilmore jforbes jsmith jwb kalev maxamillion nirik sgallagh
16:00:00 <nirik> #topic init process
16:00:13 <mprahl> .hello
16:00:13 <zodbot> mprahl: (hello <an alias, 1 argument>) -- Alias for "hellomynameis $1".
16:00:18 <mprahl> .hello mprahl
16:00:19 <zodbot> mprahl: mprahl 'Matt Prahl' <mprahl@redhat.com>
16:00:24 <sgallagh> .hello sgallagh
16:00:25 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
16:00:25 <jforbes> .hello jforbes
16:00:28 <zodbot> jforbes: jforbes 'Justin M. Forbes' <jforbes@redhat.com>
16:01:27 * nirik waits to see if we have quorum.
16:02:10 * threebean 
16:04:01 <nirik> jwb / dgilmore / jsmith ?
16:04:11 <nirik> (I know maxamillion is out today)
16:04:54 <dgilmore> hi
16:05:13 <nirik> thats 4... we need one more for quorum...
16:07:01 <nirik> so we could cancel due to lack of quorum, but there's at least the arbitrary branching which is under a very short deadline...
16:07:19 <jwb> Sorry, out today.  Have to get kids for 1/2 day
16:07:27 <jwb> On my phone right now
16:08:06 <nirik> jwb: ok, should we consider you not here for quorum purposes?
16:09:01 <jforbes> he said out today
16:09:10 <nirik> yeah.
16:09:42 <nirik> mprahl / threebean: how much would another week hurt your plans? perhaps we should discuss and try and gather more votes in ticket?
16:10:05 * threebean thinks
16:10:24 <threebean> well, our date for pagure on dist-git already got pushed by a few weeks to July 5th
16:10:31 <threebean> .. for other reasons.
16:10:42 <nirik> we also slipped a week on beta...
16:11:21 <threebean> nirik: we were going to submit an FBR to put the banner ad on pkgdb today and send something out to devel-announce.  that's the only work items that are really blocked by the fesco decision.
16:11:35 <threebean> mprahl: is that correct?  can we still advance work on other bits in the meantime?
16:11:42 <sgallagh> FBR?
16:11:46 <threebean> freeze-break-request
16:11:49 <sgallagh> Ah
16:11:55 <nirik> for fedora infrastructure.
16:12:02 <mprahl> nirik: Can we talk about it today to make sure there are no additional concerns that come up with our timeline and wiki page please?
16:12:15 <nirik> #topic #1708 - F27 System Wide Change: Arbitrary Branching
16:12:16 <nirik> .fesco 1708
16:12:16 <nirik> https://pagure.io/fesco/issue/1708
16:12:16 <dgilmore> threebean: getting it communicated now is a big +1
16:12:16 <zodbot> nirik: Issue #1708: F27 System Wide Change: Arbitrary Branching - fesco - Pagure - https://pagure.io/fesco/issue/1708
16:12:20 <mprahl> I don't think we need an answer today though
16:12:36 <nirik> mprahl / threebean: can you go over what you added in the last week?
16:12:57 <dgilmore> I am still extremely concerned over the removal of pkgdb on such a tight schedule
16:13:09 <threebean> for one, we tried to list out a more detailed timeline at the bottom of the Change here https://fedoraproject.org/wiki/Changes/ArbitraryBranching#Timeline
16:13:50 <mprahl> dgilmore: We pushed the date back to July 10th. Does that help your comfort level at all?
16:13:51 <nirik> dgilmore: yeah, it is tight for sure. ;( But less stuff to maintain is good...
16:14:13 <dgilmore> mprahl: not at all
16:14:34 <dgilmore> mprahl: I would like at least one full fedora cycle
16:14:42 <dgilmore> and for things to be phased in
16:15:37 <threebean> ok - our takeaway from last week was that the main concern from fesco was over announcement/education.
16:16:00 <jforbes> Well, that was the consensus.
16:16:01 <nirik> I thought there were reasons it couldn't be phased in...
16:16:13 <jforbes> Right
16:16:19 <threebean> nirik: yeah, not without going back to the drawing board.
16:16:53 <mprahl> nirik: We'd have to rewrite PkgDB to phase it in and even then I wouldn't really consider that phasing in unfortunately.
16:16:54 <nirik> right.
16:17:01 <threebean> (towards education, mprahl started putting together a FAQ on how to use the new bits https://fedoraproject.org/wiki/Infrastructure/WhatHappenedToPkgdb#How_do_I_give_a_user_commit_access_to_a_dist-git_repo.3F )
16:18:28 <nirik> lots of TODO's still. ;)
16:19:41 <nirik> I think it's a pretty short ramp, but I'm ok with it in general. I think you may get a bunch more feedback from posting to devel-announce... as lots of people I think have just been ignoring modules until they affect them... which this will do. ;)
16:20:09 <threebean> heh, yeah.
16:20:45 <nirik> is there any sense of progress on the pagure over dist-git work? I've not seen much going on there... is that going to happen in time?
16:21:11 <mprahl> nirik: Some of those TODOs require policy decisions which I don't feel comfortable asking people to make until FESCo approves the change.
16:21:27 * nirik nods
16:21:54 <threebean> puiterwijk's re-architecting for performance is part of it.  we "OK'd" the July 5th date with pingou.  he's been afk this week and I'm assuming it can pick back up next week.
16:22:39 <mprahl> nirik: We also had discussions with pingou and pfrields a couple weeks ago and it seems like it's one of their priorities.
16:23:02 <mprahl> and as threebean mentioned, pingou said July 5th was ok for prod.
16:23:22 <threebean> dgilmore: in the timeline section, did you see the list of releng scripts we identified to patch?
16:23:46 <dgilmore> threebean: yes, can not guarantee we do not have other things using it
16:23:58 <dgilmore> threebean: there is also a bunch of releng stuff thats in infra ansible
16:23:58 <nirik> how about you all mail devel-announce and say this is the proposed not yet approved plan for feedback... and we revisit next week? since we don't have quorum and dgilmore sounds like a -1 anyhow?
16:24:20 <dgilmore> nirik: I would say i am 0
16:24:26 <nirik> ok.
16:24:42 <dgilmore> I am extremely concerned over yanking out pkgdb
16:25:02 <dgilmore> there is a bunch of places we rely on it, some of which are likely not documented
16:25:37 <nirik> sure. we will have to be ready to fix those...
16:26:03 <dgilmore> threebean: for instance the script that syncs owners from pkgdb to koji is not mentioned
16:26:22 <dgilmore> threebean: that is responsible for adding new packages as well
16:26:24 <threebean> oh, we already have the patch for that one done, right mprahl?
16:26:39 <mprahl> Yes, I patched that script to work in the new world.
16:26:56 <mprahl> "new world" I should say
16:27:31 <threebean> dgilmore: but, good example.  there are lots of things like that scattered around.
16:27:36 <threebean> doing our best to tick them all off.
16:27:43 <threebean> heh, not irritate them - but fix them.
16:27:44 <dgilmore> threebean: releng will be the central point for dealing with breakages and we do not have anyone spare to deal with the fallout
16:28:21 <threebean> understandable.  please route things to us to fix.
16:28:48 <threebean> (maybe this is outside the scope of the meeting, but..) we could even join the releng group proper if that makes more sense.
16:28:55 <dgilmore> I strongly wouyld like more time to identify as much as we can beforehand
16:29:34 <dgilmore> have we looked at the pkgdb logs to identify clients or internal ips hitting it for things to ensure we catch the bulk of use cases?
16:29:41 <nirik> yeah, we did.
16:29:57 <nirik> thats when we found the gnome-software looking for active releases for dist upgrade thing
16:30:02 <dgilmore> threebean: it likely would make sense
16:30:10 <nirik> s/found/realized it/
16:30:27 <dgilmore> right, because they wanted us to manually maintain a file and we said no use pkgdb
16:30:57 <threebean> yeah - for the transition, we're going to auto-generate that file via cron and put it on the proxies for gnome-software
16:31:34 <nirik> whats the input data for that? or a human changes it?
16:31:57 <dgilmore> nirik: for the data it uses?
16:32:04 <nirik> yeah
16:32:10 <dgilmore> its the release state inside of pkgdb
16:32:20 <nirik> yes, now... I meant, what will it be.
16:32:33 <dgilmore> when we change the state of the release at GA time it changes automatically
16:32:43 <dgilmore> oh who knows
16:32:47 <nirik> well, no, it needs a human to switch it. ;)
16:32:56 <nirik> in the end it comes back down to fragile humans.
16:33:16 <dgilmore> it should come from something in pdc
16:33:17 <nirik> we don't (yet) have a fedmsg for 'Fedora 26 is released'
16:33:28 <dgilmore> but to date we have not correctly set up PDC at all
16:33:53 <nirik> yeah, this will make us have to.
16:34:16 <dgilmore> we also have noone outside of threebean that knows pdc
16:34:25 <threebean> well, mboddu got an introduction to it.
16:34:33 <threebean> we sat down and poked at the API together.
16:34:42 <dgilmore> threebean: he has not shown to me he understands any of it yet
16:34:46 <mprahl> mboddu was talking to ahills about it yesterday.
16:34:49 <dgilmore> but that is good to know
16:34:57 <threebean> yeah, getting there.  will take more time/work.
16:34:57 <mprahl> It seems like he is trying to learn it
16:35:17 <dgilmore> mprahl: sure, but there is a difference between learning and knowing
16:35:48 <nirik> anyhow... so would it work to do a devel-announce post today and revist next week while moving along on anything else you can without a concrete approval?
16:36:25 <nirik> only other option would be to ask folks to vote in ticket... but we have had poor luck getting people to do that over the years.
16:36:25 <threebean> yeah - that sounds good to me.
16:36:53 <threebean> mprahl and I can write it up.  nirik, dgilmore: would you guys mind proof-reading it out of band before we hit send?
16:36:57 <jforbes> I am comfortable with things moving along over the next week until we can get a vote. I don't see any concerns that have not already been raised, and are being addressed
16:37:09 <nirik> sure, happy to help look over anything...
16:37:35 * nirik is ok with the plan in general... its going to be a lot of work in a short window, but I think we can do it. :)
16:37:56 <threebean> :)
16:38:58 <nirik> #info devel-announce post will go out soon with the proposed plan and timeline
16:39:13 <nirik> #info will revisit next week
16:39:26 <nirik> thanks for being here threebean and mprahl. Appreciated.
16:39:34 <nirik> #topic Next week chair
16:39:35 <dgilmore> thanks guys
16:39:39 <nirik> anyone want to take it next week?
16:39:45 <mprahl> Thanks for your feedback :)
16:40:28 <jforbes> There aren't many of us here, I guess i will take it again
16:40:30 <threebean> (thanks!)
16:40:37 <nirik> jforbes: thanks.
16:40:42 <nirik> #info jforbes to chair next week
16:40:46 <nirik> #topic Open Floor
16:40:50 <nirik> anything for open floor?
16:41:02 <sgallagh> One item
16:41:16 <sgallagh> I just wanted to reconfirm my memory, since I couldn't find evidence one way or another:
16:41:58 <sgallagh> FESCo granted the WGs the privilege to declare certain tickets blockers irrespective of criteria if they deem it sufficiently important. Is that correct?
16:42:30 <dgilmore> I do not remeber that
16:42:36 <dgilmore> but does not mean it is not so
16:44:06 <sgallagh> I feel like that was one of our early decisions about the WGs as far as granting them autonomy
16:44:22 <jforbes> Would have bene before my time, maybe search the logs?
16:44:23 <sgallagh> That within their scope, they had that leeway to hold up the release if it was important enough for their Edition
16:45:02 <nirik> I can't recall either...
16:45:13 <sgallagh> jforbes: too many logs to search :-/
16:45:52 <dgilmore> sgallagh: https://pagure.io/fesco/issue/1202
16:46:06 <dgilmore> that is the only thing I am getting when using google
16:46:58 <sgallagh> Yeah, I'm not coming up with it either, but I swear we discussed it. I have no proof to back that up
16:47:52 <sgallagh> It came up recently over a disagreement between the Workstation WG and the blivet-GUI folks.
16:47:57 <dgilmore> sgallagh: I am inclined to say we did not agree to that
16:48:49 <sgallagh> Essentially they want blivet-GUI to repackage so it doesn't show up as a default app in GNOME because it doesn't meet their guidelines for inclusion.
16:48:51 <nirik> well, we could look more, and failing that you can add it to next weeks agenda?
16:49:03 <sgallagh> But it gets installed via the live installation
16:49:16 <sgallagh> And they have their desktop file packaged with their main functionality
16:49:28 <dgilmore> sgallagh: I suggest you file a ticket and we can look at it next week
16:49:32 * sgallagh nods
16:49:59 <dgilmore> we can discuss in ticket if need ba
16:50:03 <dgilmore> need be
16:50:09 <sgallagh> Sure
16:50:14 <dgilmore> since hopefully by next week the release is go
16:50:22 <sgallagh> I mostly brought it up because I was hoping someone else would just remind me where that decision was :)
16:50:41 <sgallagh> I think they only wanted it as a Final blocker, so it's not quite that urgent
16:50:46 <dgilmore> sgallagh: I honestly do not remeber ever making a decision like that
16:51:06 <sgallagh> dgilmore: My memory could be playing tricks on me or conflating it with a similar decision
16:51:25 <nirik> anyhow, if nothing else will close out in a minute...
16:51:44 <sgallagh> WFM
16:52:33 <nirik> thanks for coming everyone.
16:52:37 <nirik> #endmeeting