18:00:02 <smooge> #startmeeting EPEL (2019-02-27)
18:00:02 <zodbot> Meeting started Wed Feb 27 18:00:02 2019 UTC.
18:00:02 <zodbot> This meeting is logged and archived in a public location.
18:00:02 <zodbot> The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:02 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:02 <zodbot> The meeting name has been set to 'epel_(2019-02-27)'
18:00:02 <smooge> #meetingname epel
18:00:02 <zodbot> The meeting name has been set to 'epel'
18:00:02 <smooge> #topic Chair and Introductions
18:00:02 <smooge> #chair avij bstinson Evolution nirik smooge pgreco tdawson
18:00:02 <zodbot> Current chairs: Evolution avij bstinson nirik pgreco smooge tdawson
18:00:12 <smooge> hello everyone
18:00:15 <bstinson> hi all
18:00:17 <kanarip> good morning
18:00:21 <smooge> nirik, is out for move
18:00:31 <pgreco> hello
18:00:43 <tdawson> hello
18:00:47 <Evolution> o/
18:01:08 <smooge> #topic Agenda
18:01:08 <smooge> #info EPEL-7 Proposal: Release_based_package_lifetimes
18:01:08 <smooge> #info EPEL-7 Proposal: Minor_release_based_composes
18:01:08 <smooge> #info EPEL-8 Tentative Proposal
18:01:08 <smooge> #info Python36 Flag Day
18:01:23 <smooge> Anything else for the agenda today?
18:02:58 <smooge> #topic EPEL-7 Proposal: Release_based_package_lifetimes
18:02:58 <smooge> #info https://fedoraproject.org/wiki/EPEL/Changes/Release_based_package_lifetimes
18:03:47 <smooge> OK anything people want to discuss or fix in this?
18:04:45 <bstinson> this process makes sense to me
18:05:28 <bstinson> should we define who the "EPEL release manager" is in our docs?
18:05:33 <pgreco> doesn't look like it ;)
18:05:45 <tdawson> To me it makes sense, and it goes well with the other doc we will be discussing.
18:06:34 <smooge> bstinson, we should define that person. I left it vague because I wasn't sure if that needed to be assigned by FESCO
18:07:23 <bstinson> #proposal Accept the Released Based Package Lifetimes proposal as written, and notify FESCO, FPC, and the EPEL lists of our action
18:07:43 <smooge> I second that proposal so +1
18:07:49 <pgreco> +1
18:07:58 <bstinson> +1
18:08:12 <tdawson> +1
18:08:19 <kanarip> in case i get a vote: +1
18:08:49 <smooge> well for the committee currently I need Evolution's vote :)
18:09:01 <kanarip> that's what i thought
18:09:17 <smooge> but your interest is noted
18:10:26 <Evolution> +1
18:10:50 <Evolution> I'm here. just buried in budget/email fun
18:11:07 <smooge> OK with nirik not here, I am going to say
18:11:31 <Evolution> smooge: I have a hand-puppet you can use for nirik's vote if you want.
18:11:37 <Evolution> sharky can say....
18:11:39 <Evolution> :-P
18:12:10 <smooge> #approved Accept the Released Based Package Lifetimes proposal as written, and notify FESCO, FPC, and the EPEL lists of our action (+5, 1 abstain, 0 -, +1 from audience)
18:12:22 <smooge> #topic EPEL-7 Proposal: Minor_release_based_composes
18:12:22 <smooge> #info https://fedoraproject.org/wiki/EPEL/Changes/Minor_release_based_composes
18:12:34 <smooge> #approve Accept the Released Based Package Lifetimes proposal as written, and notify FESCO, FPC, and the EPEL lists of our action (+5, 1 abstain, 0 -, +1 from audience)
18:12:49 <smooge> I can never remember which one gets flagged in the notes
18:13:41 <smooge> ok so this one is probably going to need the Mythical Release Engineer time the most
18:14:28 <pgreco> I like the concept of this, don't know that it needs to be implemented
18:15:43 <smooge> so the major reason I see needing this is that I believe it will be needed for RHEL-8
18:15:47 <kanarip> i find there's a typo in the first symlink example
18:16:24 <kanarip> namely: /pub/epel/releases/7 -> /pub/epel/updates/7.06.201903
18:16:31 <smooge> oops
18:17:11 <kanarip> i also believe this will be essential in rhel8 space, and of particular interest to .z stream consumers, of whom i'm encountering more and more
18:17:16 <kanarip> which*?
18:17:26 <Evolution> I worry about changing epel7, given that people have had a number of years to get comfy with it
18:17:30 <Evolution> I'm fine with doing the change for 8.
18:17:58 <kanarip> nothing functionally changes for epel7 though, so long as you keep things default, right?
18:18:16 <bstinson> i don't think the user-facing experience changes, except that yum downgrade now works
18:18:20 <kanarip> just like a centos 'yum update' would get you whatever is then called '7' by symlink
18:18:26 <smooge> Evolution, the problem was that we still have 5 more years of support epel7 and the 'well for this release we do X and that release does Y' ends up causing pain in the past
18:18:32 <pgreco> Evolution, if we do this right, the user shouldn't even notice
18:18:37 <Evolution> fair.
18:18:59 <smooge> as in 'oh we can't update koji/bodhi/etc because EPEL-old still needs to be maintained for N more years'
18:19:08 <smooge> so I am hoping to make it so the users shouldn't notice
18:19:19 <smooge> and we can make it so we don't have as many special cases
18:20:07 <smooge> many of this was on the books to be tried when EPEL-7 was planned but we couldn't at the time because it would break EPEL-5 :)
18:20:34 <smooge> the systems seem more robust so we should not break EPEL-6 if this works
18:20:37 <kanarip> i would see a potential to replace an 'epel-release-8-1.el8' with an 'epel-release-z-8.1-8-1.el8', despite the fact that may just not happen
18:20:56 <kanarip> (someone would have to type 'dnf shell' perhaps)
18:20:56 <smooge> so my plan to see if it works is that we will test this in /pub/alt/epel first
18:21:03 <tdawson> I know that both as a user and as a maintainer, it's a bit of a pain to have to keep the old stuff when I and my users really needed the newer stuff.  I feel this will address that issue nicely.
18:21:42 <tdawson> So, while I know there are details, (like how to do it, and how to get the old stuff to the users) I really like this plan.
18:21:47 <smooge> so I would like to get the engineering in place, make the rollouts happen in /pub/alt/epel.. get as much feedback as we can and then make it happen in /pub/epel
18:22:42 <smooge> I am completely happy to throw this proposal out if our testing in /pub/alt/epel comes up a burning coalmine
18:22:43 <kanarip> can this obsolete/purge from mirror locations versions of EPEL for which RHEL has no z-stream support anymore?
18:23:03 <kanarip> or does this solely follow fedora release cycles?
18:23:48 <kanarip> (as opposed to rhel minor releases)
18:24:06 <smooge> I don't know yet. I would prefer rhel minor releases but it may be FRC's
18:24:35 <kanarip> i would also prefer rhel minor releases
18:24:57 <kanarip> on another point, the examples seem to omit current users expect Stuff/ to be available, and Moduler to be added as the new thing
18:25:51 <smooge> oh foo
18:26:04 <smooge> modular shouldn't be in the 7 proposal at all
18:26:14 <smooge> as you can tell this was originally an 8 one
18:26:18 <smooge> I will fix that also
18:26:24 <kanarip> yup
18:27:03 <smooge> no.. this covers all releases
18:27:09 <smooge> except 6
18:27:46 <kanarip> but then requires an update to epel-release, correct?
18:28:20 <kanarip> and moreover requires static/homegrown configs to be updated...?
18:29:19 <smooge> kanarip, the epel-release will have to be updated. We had do that in the past when we changed composers
18:29:29 <smooge> and the layout of directories changed
18:29:54 <kanarip> and everyone with homegrown configs will have to catch up to the new black
18:30:08 <smooge> I am hoping to deal with some of the problems with symlinks but homegrown will have to catch up
18:30:33 <kanarip> there's no /pub/epel/{releases,updates}/7.06.201903/$basearch/repodata/ ... and even if there were, existing configs would be assisted by symlink only to one of the two
18:31:14 <kanarip> so there would be a reason to retain /pub/epel/7/$basearch -> /pub/epel/releases/7.06.201903/Stuff/$basearch/
18:31:28 <kanarip> it'll make yum update to the next epel-release actually function
18:32:19 <smooge> so general question.. is that a policy problem or a detail problem?
18:32:26 <kanarip> detail
18:32:27 <bstinson> to match the existing functionality, it would need to link to /pub/epel/updates/7.06... correct?
18:32:30 <kanarip> perhaps offline is better
18:32:48 <kanarip> bshephar, updates is empty at the beginning of 7.06.201903's release
18:33:17 <kanarip> wrong b-tab completion, bstinson ^^
18:34:15 <bstinson> for the initial bootstrap, it wouldn't be empty. releases,updates would each contain a copy of the updated epel-release
18:34:41 <bstinson> if you link to releases/7.06 there's no transition period. and folks on the existing setup don't get updates to currently installed packages
18:34:59 <smooge> yeah this was how nirik thought it through last meeting
18:35:01 <kanarip> but /pub/epel/7/$basearch points nowhere, and it can only link to one place
18:35:23 <smooge> so for the initial bootstrap we would have it point to updates
18:35:27 <kanarip> either stuff or modules, and either releases or updates for either of those
18:35:34 <smooge> after the first release we point it to releases
18:35:46 <smooge> second release
18:35:52 <kanarip> right
18:35:56 <kanarip> ass... covered
18:36:28 <smooge> so first release on day 0 we have it point to updates with updates containing everything that EPEL-8 has on day 0. Then after that it gets updated until the next release
18:36:47 <kanarip> epel 7*
18:36:52 <smooge> EPEL-7
18:37:05 <smooge> I typed 5 the first tie
18:37:07 <kanarip> epel 8 can dive straight in ;-)
18:38:16 <smooge> ok
18:39:10 <smooge> #proposal  Accept the Minor Release Based Composes proposal as written, and notify FESCO, FPC, and the EPEL lists of our action. Update/fix as things flesh out in reality
18:39:59 <bstinson> +1, stongly in favor
18:40:09 <tdawson> +1
18:40:19 <kanarip> +1 here too in case you wanted to type 'from audience' again
18:40:26 <smooge> +1
18:40:41 <pgreco> +1
18:40:49 <smooge> Evolution down to you
18:41:06 <kanarip> smooge, side-detail -- mirrormanager updates for this
18:41:15 <smooge> yep.. so many
18:41:21 <kanarip> not rel-eng iirc?
18:41:26 <smooge> they will appear as releases
18:41:37 <kanarip> you got it covered i'm sure
18:41:41 <smooge> hahahahahahahahaa
18:41:56 <Evolution> t+1
18:41:57 <smooge> sorry where was i ? insane laughter
18:41:58 <Evolution> bah
18:41:59 <Evolution> dammit.
18:42:01 <Evolution> +1
18:42:54 <smooge> #approve  Accept the Minor Release Based Composes proposal as written, and notify FESCO, FPC, and the EPEL lists of our action. Update/fix as things flesh out in reality (+5, 1 abstain, 0 -, +1 from audience)
18:43:17 <smooge> ok yay.. we did real policy work..
18:43:31 <kanarip> now for the meat of the meeting
18:43:37 <smooge> #topic EPEL-8 Tentative Proposal:
18:43:37 <smooge> #info https://fedoraproject.org/wiki/Infrastructure_2020/EPEL-8
18:44:22 <smooge> ok so the above was my plan for getting time/effort/work on the Fedora Infrastructure calendar so EPEL was not done adhoc
18:44:53 <smooge> it needs more work and feedback via email as needed
18:45:14 <Evolution> I'd like to volunteer tdawson to help scope some of this.
18:46:14 <tdawson> I'm fine with that, just know that I don't know the fedora infrastruture as well as others.
18:47:18 <Evolution> that's where I point at smooge and nirik if he were here.
18:47:30 <smooge> cool. I would appreciate it. One of the policy items we will need to work out is how we seed it
18:47:37 <kanarip> i know some of what goes on there too
18:47:47 <tdawson> But it looks like bstinson is already on the list
18:48:10 <smooge> tdawson, that was mostly me volunteering people
18:48:18 <smooge> and more than one person can fill out those role
18:48:20 <bstinson> it never hurts to have extra names
18:48:35 <smooge> Brian was also my professional heckler at my talk on this..
18:48:48 <tdawson> :)
18:49:22 <smooge> tdawson, how does your friday look this week?
18:49:55 <tdawson> smooge: Have an hour meeting at the very beginning of my day, rest of day is free
18:50:07 <kanarip> i have 1203 out of 22xx packages rebuilt against one another without help from external repositories/dependencies
18:50:09 <smooge> ok you are central us timezone
18:50:36 <tdawson> smooge: That's the time zone I work in, yes
18:51:02 <tdawson> (not the time zone I live in)
18:51:08 * Evolution looks at the map, and looks at tdawson
18:51:14 <Evolution> heh
18:51:17 <smooge> so does this 10 am your time 1600 UTC sound ogod
18:51:20 <kanarip> well, he definitely isn't a timezone
18:51:43 <smooge> ugh sorry I missed that
18:51:55 <tdawson> I'm living in Spokane Washington (same time zone as L.A.), but I work in the Central Time zone ... I just get up really early.
18:52:17 <smooge> ooooh I thought you were inside the  Fermi collider still
18:52:51 <Evolution> no reason to stay once he got his super powers
18:52:59 <tdawson> smooge: Well, I tried to stay out of the collider, at least when it's running. :)  but it's been a year since I moved out here.
18:54:12 <tdawson> Feel free to pretend I'm over there, it makes things easier.
18:54:28 <smooge> ok so let us meet up at 1600 UTC Friday in #epel and go over where we all are at... aka a pre-meeting-meeting as I think we will need to know that before we can look at what we will break koji+bodhi+etc for EPEL-8-beta
18:55:02 <smooge> tdawson, Evolution stole my joke about superpowers.. with your secret lair in the center of the collider
18:55:11 <Evolution> :-P
18:55:11 <smooge> ZapCave
18:55:17 <kanarip> i'm going to participate because i'll have to tell SmootherFrOgZ ;-)
18:55:31 <tdawson> *laughs*
18:55:32 <Evolution> smooge: I'm a manager. ruining other peoples' fun is part of the job r
18:55:34 <Evolution> -r
18:55:37 * bstinson adds to the calendar
18:55:48 <pgreco> Evolution, it is the best part of that job
18:55:48 <smooge> yeah.. rpmfusion and such will be impacted
18:56:01 <kanarip> i'll blog the isv perspective ;-)
18:56:29 <smooge> ok I think that is it for this meeting.. I am going to move ot open floor as I am starting to type like johnny
18:56:37 <smooge> #topic open flood
18:56:42 <kanarip> hahaha
18:56:54 <kanarip> involuntary property baptism?
18:57:20 <smooge> yeah. thank you all for your time this week. I hope to have a report on EPEL-8 next week.
18:57:37 <smooge> I would also like to work out a way to do alternative arches in this
18:57:55 <SmootherFrOgZ> kanarip: ish
18:58:12 <kanarip> SmootherFrOgZ, probablish
18:58:13 <smooge> so we can have that Risc-V EPEL-8 we all want
18:58:35 <SmootherFrOgZ> heh
18:58:41 <Evolution> the 8 users world-wide will be very happy about that
18:58:59 <smooge> if there isnt;' anytihng else and my typing is worseniz wthout cofeef
18:59:20 <Evolution> smooge: nothing further. thanks for driving
18:59:25 <smooge> [that was me not fixing any of my typing which is why it is so slooow today]
18:59:34 <tdawson> nothing from me
19:00:00 <smooge> #info There will be an EPEL-8 pre-planning-planning meeting on Friday 2019-03-01 at 1600 UTC in #epel
19:00:05 <smooge> #endmeeting