16:00:36 <mboddu> #startmeeting RELENG (2019-11-07)
16:00:36 <zodbot> Meeting started Wed Nov  6 16:00:36 2019 UTC.
16:00:36 <zodbot> This meeting is logged and archived in a public location.
16:00:36 <zodbot> The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:36 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:36 <zodbot> The meeting name has been set to 'releng_(2019-11-07)'
16:00:36 <mboddu> #meetingname releng
16:00:37 <mboddu> #chair nirik sharkcz pbrobinson pingou mboddu dustymabe ksinny jednorozec
16:00:37 <mboddu> #topic init process
16:00:37 <zodbot> The meeting name has been set to 'releng'
16:00:37 <zodbot> Current chairs: dustymabe jednorozec ksinny mboddu nirik pbrobinson pingou sharkcz
16:01:57 * mboddu gets something to drink in the mean time
16:02:02 <nirik> morning
16:03:44 <mboddu> morning nirik
16:03:58 <mboddu> More reasons to hate DST, I guess :)
16:04:48 <nirik> as if we needed more
16:05:02 <mboddu> Haha :)
16:05:13 <mboddu> So, lets get started
16:05:36 <mboddu> #topic EPEL 8
16:05:52 <mboddu> merlinm: If possible can you give us a quick update?
16:09:13 <merlinm> sorry, what's the question?
16:09:42 <merlinm> I think I see.
16:10:02 <merlinm> Modules are building in staging for el8 and el8_playground!
16:10:23 <nirik> hurray!
16:10:37 <merlinm> I'm trying to figure out bodhi/pungi configs to do something with them.
16:10:40 <nirik> I guess next we need a compose for them?
16:10:41 <nirik> yeah
16:10:48 <mboddu> merlinm: I can help you with that
16:10:53 <merlinm> But I'm not overly familiar with those configs.
16:10:56 <merlinm> Awesome!
16:11:19 <mboddu> Okay, can you create a ticket with releng, just like you did for epel module tags to track this work?
16:11:24 <mboddu> merlinm: ^
16:11:33 <merlinm> mboddu: will do
16:11:42 <mboddu> merlinm++
16:11:51 <mboddu> Thanks merlinm for your help on this
16:12:06 <mboddu> #info merlinm is able to build epel8 and epel8-playground builds in stg
16:12:31 <mboddu> #info Next step remaining is composes and bodhi updates
16:13:48 <mboddu> merlinm: epel-playground modules config will be part of the nightly epel8-playground rpm compose config
16:14:02 <mboddu> We need to update it
16:14:20 <mboddu> And that leaves bodhi config and sync script updates
16:14:28 <mboddu> And obviously adding the bodhi release
16:14:32 <mboddu> Am I missing anything?
16:15:48 <merlinm> I think that's all. But then again, this is unfamiliar territory for me. :)
16:16:18 <mboddu> merlinm: Understandable, which is why we are here to help :)
16:16:56 <mboddu> nirik: So, that leaves the big elephant in the room, rhel 8.1
16:17:08 <nirik> what about it? ;)
16:17:59 <nirik> we should just be able to sync it no?
16:18:30 <mboddu> nirik: I guess so, but, just wanted to check
16:19:21 <nirik> yeah, should be fine(tm)!
16:20:29 <mboddu> nirik: Dont we have to update any rhn sync scripts?
16:20:34 * dustymabe waves
16:20:48 <nirik> yes,, talking with smooge about it in fedora-noc now. ;)
16:21:55 * mboddu waves back at dustymabe
16:22:57 <nirik> it should be s/8.0/8.1/ in some configs, but we will see.
16:23:52 <mboddu> Yeah, seems like some flags got changed and rpm format changes
16:24:57 <mboddu> #info We need to update the rhn sync scripts and some configs, but we are waiting on mass reboots and there are some things consider like adding modularity flags and rpm format changes
16:25:48 <mboddu> Anything else on this topic?
16:26:13 * nirik doesn't have anything
16:27:19 <mboddu> Okay, moving on
16:28:53 <merlinm> mboddu: FYI, I created two tickets for the follow-on modular tasks. https://pagure.io/releng/issue/8968 for el8, https://pagure.io/releng/issue/8967 for playground
16:29:20 <mboddu> merlinm: Thanks
16:29:25 <mboddu> #topic #8811 method for copying in ostree content to our ostree repos
16:29:30 <mboddu> #link https://pagure.io/releng/issue/8811
16:29:35 <mboddu> dustymabe: You are up :)
16:32:05 <dustymabe> mboddu: hi
16:32:26 <dustymabe> mboddu: this one I think is on me to get with nirik and see if we can make progress
16:32:31 <dustymabe> nirik: is that right?
16:33:01 <nirik> yeah, I think so... I have currently been swamped.
16:33:08 <nirik> could we try early next week?
16:33:28 <nirik> (if we pick a time/date it's more likely to happen I think)
16:33:42 <dustymabe> nirik: I have a f2f next week so I'll be on the road. I think it would only work if it was first thing in the morning on Monday
16:33:53 <dustymabe> but that would mean you'd be getting up really early
16:34:00 <dustymabe> nirik: we can punt til the following week maybe
16:34:07 <dustymabe> but would like to make progress soonish
16:34:15 <dustymabe> sound good?
16:35:14 <nirik> sure.
16:35:19 <nirik> whatever works.
16:35:32 <nirik> following week might be good.
16:35:38 <mboddu> #info nirik will work with dustymabe in the next couple of weeks to get this done.
16:35:40 <nirik> perhaps we can schedule once you are back?
16:36:10 <dustymabe> can we discuss the other ticket real quick
16:36:28 <mboddu> dustymabe: Which one?
16:36:52 <dustymabe> https://pagure.io/releng/issue/8421#comment-609609
16:36:57 <dustymabe> sorry was digging up link
16:38:02 * nirik reads
16:39:16 <nirik> yeah, I think we can grant that... but the 'user' needs to login first I think?
16:39:35 <dustymabe> nirik: the user is running already
16:39:47 <dustymabe> it's what is doing the tagging for the coreos-pool tag
16:39:50 <nirik> odd:
16:39:53 <nirik> ~ stg-koji list-permissions --user=coreos-koji-tagger/coreos-koji-tagger.phx2.fedoraproject.org@STG.FEDORAPROJECT.ORG
16:39:53 <nirik> User coreos-koji-tagger/coreos-koji-tagger.phx2.fedoraproject.org@STG.FEDORAPROJECT.ORG does not exist
16:40:09 <dustymabe> nirik: I took down the project in stage
16:40:14 <dustymabe> since I'm not testing it right now
16:40:19 <nirik> ok
16:40:23 <dustymabe> I can deploy it to stage if you'd like
16:41:00 <nirik> ah, it's coreos-koji-tagger/coreos-koji-tagger.fedoraproject.org@FEDORAPROJECT.ORG in prod
16:41:04 <nirik> (no .phx2)
16:41:48 <dustymabe> so would you like me to deploy it to stg ?
16:42:14 <nirik> naw.
16:42:34 <nirik> well, doesn't matter much aside from we will need to grant the perm in stg for testing when you next need to test it.
16:42:41 <nirik> ok, set it in prod
16:42:56 <nirik> ~ koji list-permissions --user=coreos-koji-tagger/coreos-koji-tagger.fedoraproject.org@FEDORAPROJECT.ORG
16:42:56 <nirik> coreos-continuous
16:42:56 <nirik> dist-repo
16:43:02 <mboddu> "koji grant-permission dist-repo coreos-koji-tagger/coreos-koji-tagger.fedoraproject.org@FEDORAPROJECT.ORG" should do it
16:43:09 <mboddu> Well, nirik got to it
16:43:34 <nirik> yep. that. ;) Should we close that ticket then?
16:43:59 <dustymabe> nirik: we will be unblocked for now if that is working
16:44:12 <dustymabe> but longer term we'd like to give it to anyone in that group
16:44:22 <dustymabe> https://pagure.io/koji/issue/1660
16:44:58 <nirik> right, we need that implemented first
16:45:04 <mboddu> dustymabe: Yeah, we have to wait for koji to have that fixed
16:45:55 <mboddu> dustymabe: Can you close the releng ticket, once you verify its working?
16:47:16 <dustymabe> mboddu: yep
16:47:18 <dustymabe> thanks
16:48:21 <mboddu> Thanks dustymabe
16:49:41 <mboddu> Lets dig at one of the old tickets - found while pagure grooming
16:49:49 <mboddu> #topic #8026 Packages are not retired entirely
16:49:54 <mboddu> #link https://pagure.io/releng/issue/8026
16:50:24 <mboddu> So, the discussion is basically whether we should remove every maintainer when the package gets retired or not?
16:52:04 * nirik isn't sure, would have to real this in detail. ;(
16:52:28 * mboddu not sure either
16:52:35 <nirik> I'm not sure what harm having maintainers on a retired package is?
16:52:50 <mboddu> Esp now that people cannot push with pdc hook
16:52:52 <nirik> I guess you could still get bugs on it.
16:53:01 <mboddu> Yes
16:53:05 <nirik> but we should fix that in bugzilla.
16:54:02 <nirik> it might mess up stats... ie, how many packges you maintain or something.
16:54:10 <mboddu> Right
16:54:49 <nirik> but it also gives history... that you _were_ the maintainer...
16:55:26 <mboddu> I guess, removing everyone seems like a good thing to do, but that means, maintainer B cant pick it up just because maintainer A decided to retire it
16:55:36 <nirik> I guess I'd say might be a discussion for devel, but I'd hate to bring it up by itself instead of a full overhaul plan of some kind
16:55:48 <nirik> yeah.
16:57:14 <mboddu> Yeah, not a good idea to post it on devel@ without couple of options, or else it will be a rant of different things from different people
16:59:02 <mboddu> Or else add a flag to fedpkg retire like --all which will remove everyone as the maintainers
16:59:26 <mboddu> And assign it to orphan
17:00:14 <nirik> I don't think it should be another option...
17:00:31 <nirik> we should either always do it, or not... but I am not sure whats the better option.
17:00:37 <nirik> will need to ponder on it.
17:02:36 <mizdebsk> removing maintainers would lead to situation where there is no one to grant access to the package for new maintainers, eg. when new arbitrary branch is requested, resulting in more manual work for releng
17:03:20 <mizdebsk> eg. someone retires package, all people are removed, then someone else comes and asks for arbitrary branch (for a module), but no one has commit/admin access
17:03:34 <mboddu> mizdebsk: Right, there is no good option here
17:03:46 <mboddu> Well, lets punt it for now
17:03:59 <mboddu> We can come back to it
17:04:22 <mboddu> And when pagure supports branch level acls, we can find a better option
17:04:31 <mboddu> Anything else?
17:05:42 <nirik> I have some short info on upgrades if you want...
17:05:47 <mboddu> nirik: Sure, go ahead
17:05:53 <mboddu> #topic Open Floor
17:06:02 <nirik> I'm moving stg builders to f31 (not all done yet).
17:06:37 <nirik> I moved stg koji to 1.19.0... it hit a nasty bug (boo), but there was already a fix upstream (yea!). Seems that have stablized it.
17:06:53 * mboddu guesses nirik is having a ton of fun ;)
17:07:02 <nirik> I haven't moved prod yet, possibly later in the week or next week
17:07:11 <mboddu> Sounds good
17:07:15 <nirik> and prod builders will be a slower process.
17:07:24 <nirik> but I'd like to get them all up to 31
17:07:32 <nirik> rawhide and branched composers are on 31 now.
17:07:40 <nirik> That broke the retirement script
17:08:41 <mboddu> #info nirik is upgrading all the stg builders to f31 and updated koji to 1.19.0 which had a bug and fixed it
17:08:49 <mboddu> nirik: I fixed the unretirement script
17:09:04 <nirik> yeah, was just going to say... then mboddu fixed it. :)
17:09:53 <mboddu> #info nirik is planning to update prod builders later in the week or the next week
17:10:18 <mboddu> #info rawhide and branched compose boxes are moved to f31 which broke block retired script and mboddu fixed it
17:10:26 <mboddu> Okay, I guess thats it
17:10:32 <nirik> yep
17:10:32 <mboddu> Sorry, it wont over by 10 min
17:10:40 <mboddu> Thanks everyone for joining
17:10:43 <mboddu> #endmeeting