20:00:25 <tdawson> #startmeeting EPEL (2021-07-28)
20:00:25 <zodbot> Meeting started Wed Jul 28 20:00:25 2021 UTC.
20:00:25 <zodbot> This meeting is logged and archived in a public location.
20:00:25 <zodbot> The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:25 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:25 <zodbot> The meeting name has been set to 'epel_(2021-07-28)'
20:00:25 <tdawson> #meetingname epel
20:00:25 <zodbot> The meeting name has been set to 'epel'
20:00:25 <tdawson> #chair nirik tdawson bstinson pgreco carlwgeorge michel dcavalca
20:00:25 <zodbot> Current chairs: bstinson carlwgeorge dcavalca michel nirik pgreco tdawson
20:00:25 <tdawson> #topic aloha
20:00:44 <nirik> morning
20:00:51 <tdawson> Hi nirik
20:01:08 <carlwgeorge> .hi
20:01:08 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
20:01:34 <tdawson> Hi carlwgeorge
20:01:35 <pgreco> hello!
20:02:10 <tdawson> Hi pgreco
20:03:48 <tdawson> michel already said he wouldn't be here.  I hope that dcavalca can make it
20:05:40 <tdawson> We're a little sparse today, but that's ok I guess.
20:05:46 <tdawson> #topic Old Business
20:05:53 <pgreco> I have a follow up on the tor thing for open floor
20:06:00 <tdawson> Cool
20:06:23 <tdawson> Should we jump into the bodhi thing, or wait a bit hoping others show up?
20:06:53 <tdawson> I'm going to wait a bit.
20:07:11 <tdawson> carlwgeorge how is epel-next coming along?
20:07:19 <tdawson> specifically for epel9-next
20:07:27 <carlwgeorge> https://pagure.io/releng/issue/10235
20:07:47 <carlwgeorge> no progress yet, but i've requested the c9s buildroot get mirrored so we can get started
20:08:44 <tdawson> It's the first step ... nice
20:09:21 <tdawson> I guess until that happens, there isn't really much to do.
20:09:39 <nirik> I'll try and get it this week... it's been crazy tho. :)
20:09:57 <tdawson> Thanks nirik
20:10:14 <tdawson> Anything else on next ?
20:10:49 <carlwgeorge> nope
20:11:22 <tdawson> Do we have anything for the EPEL-Packaging SIG?
20:11:50 <tdawson> michel and Davide usually comment on it, but I don't think either of them are here.
20:12:26 <tdawson> OK, then moving on.
20:13:02 <tdawson> Quick note on documentation, I haven't heard anything from Petr since his last email ... I don't know if it's good or bad.
20:13:22 <tdawson> I have heard from Adam, who let me know the steps to take to get things started.
20:13:44 <tdawson> So I'll give Petr another week, and if he hasn't replied, I'll start working through Adams steps.
20:13:56 <pgreco> sounds good
20:14:28 <tdawson> So, hopefully in a couple weeks we at least have a skeleton, either done by me or Petr.  And a way to move forward.
20:15:17 <tdawson> And with that, the bodhi stuff is what I have next.
20:15:41 <tdawson> nirik I believe you had concerns about changing it
20:15:57 <nirik> well, I don't think lower than 7 days is at all good.
20:16:08 <nirik> it takes time to mirror out, for people to update, etc.
20:16:40 <tdawson> So your concern was more about doing 3 days for epel-next?
20:17:34 <nirik> well, that and that epel is supposed to add on to a more stable product, so more time in testing is why it was 14 days in the first place.
20:18:00 <nirik> but if everyone wants to move to 7 I won't stand in the way... ;)
20:18:41 <tdawson> Understood
20:18:46 <nirik> in some kind of crazy ideal world we would have data and be able to tell if we change it that there are more bugs or something, but I don't think we have anything like that
20:19:21 <carlwgeorge> my vote is 7 days for both epel and epel-next
20:19:50 <tdawson> I'm trying my best bodhi query stuff, and having a hard time.
20:19:59 <pgreco> I liked the idea of 3 for next, but I get that it may have other complications
20:20:10 <pgreco> so 7 for both is the next best thing I guess
20:20:15 <nirik> do note that some people only ever post negative karma...
20:20:18 <rsc> (I also would prefer 7 days)
20:20:31 <tdawson> I'm good for 7 for both
20:20:32 <nirik> ie, if there's no problems they see, they just ignore it... and you see no feedback at all
20:21:05 <carlwgeorge> guilty as charged
20:21:10 <tdawson> But I'm going to try to get some sort of data, showing how often anyone has interacted (karma) after 7 days.
20:21:17 <pgreco> Yeah, I normally only do positive karma when I'm in a hurry to get it released
20:21:39 <pgreco> but it is just my vote most of the time
20:21:40 <nirik> I don't know if fedora-easy-karma works anymore or works with epel at all.
20:23:48 <tdawson> Just incase people know, it's currently 4 votes for 7 days, 1 tenative vote for 14 days
20:24:03 <tdawson> /incase/so/
20:24:37 * Eighth_Doctor waves
20:24:40 <Eighth_Doctor> .hello ngompa
20:24:41 <zodbot> Eighth_Doctor: ngompa 'Neal Gompa' <ngompa13@gmail.com>
20:24:49 <tdawson> Hi Eighth_Doctor
20:24:56 <nirik> hey Eighth_Doctor
20:25:02 <Eighth_Doctor> I'd prefer 3 for next and 7 for regular
20:25:28 <Eighth_Doctor> frankly, unless I go and poke people, I rarely get any karma for anything I submit
20:25:53 <nirik> but that doesn't mean it didn't get any testing. :)
20:26:08 * Eighth_Doctor shrugs
20:26:14 <tdawson> Eighth_Doctor: Would you be ok for 7 days for next?
20:26:35 <Eighth_Doctor> tdawson: ehh, I guess? I'd really rather the loop be shorter
20:26:55 <Eighth_Doctor> I'd want a compelling reason to say it should be 7 instead of 3 for -next
20:27:12 <nirik> well, it sometimes takes a day or two to sync to all mirrors... that doesn't leave much window for people to test.
20:27:29 <carlwgeorge> yeah 3 cuts it really close
20:27:41 <Eighth_Doctor> well then, what about 5?
20:27:49 <Eighth_Doctor> I want -next to take less time than epel
20:28:19 <carlwgeorge> 5 and 7 are pretty close, and i like the idea of having them match
20:28:22 <nirik> we don't have any epel stats here, but you can see for fedora: https://admin.fedoraproject.org/mirrormanager/propagation
20:29:01 <Eighth_Doctor> so the overwhelming majority is synced within a day
20:29:43 <Eighth_Doctor> so I don't buy the argument as much that it's not enough time
20:29:45 <nirik> yeah...
20:30:16 <nirik> it used to be much longer... so I'm ok with 3 days after looking at that... but it's still a pretty short test window...
20:30:20 <Eighth_Doctor> and I know from personal experience that if your mirror is 3 days old, the crawler disables it from the mirror network
20:30:28 <nirik> yep.
20:31:04 <Eighth_Doctor> so I am not convinced that -next shouldn't be 3
20:31:18 <Eighth_Doctor> but 7 for both is better than the current situation (14)
20:31:28 <Eighth_Doctor> but I do really think that we should have -next on 3
20:31:46 <Eighth_Doctor> because I really don't think people are actually testing
20:32:01 <nirik> well, we can't know for sure...
20:32:06 <Eighth_Doctor> and for next, they especially are unlikely to test until after it lands anyway
20:32:29 <tdawson> I'm good for 3 on -next, 7 on regular.
20:32:35 <nirik> I imagine there's lots of people with epel-testing enabled who don't add any karma.... just like I know there's lots of fedora people doing that.
20:33:21 <Eighth_Doctor> then this doesn't change anything for them
20:33:27 <carlwgeorge> my fear is that people will use the fewer testing days for epel-next as an argument for why centos stream is unusable
20:33:30 <Eighth_Doctor> stuff gets added to testing within 24 hours of submission
20:33:55 <nirik> well, I should say only reporting bugs/negative karma...
20:34:41 <Eighth_Doctor> carlwgeorge: that's a no-win situation anyway
20:34:44 <nirik> carlwgeorge: well, this is epel-next tho right? we don't promise anything?
20:35:12 <carlwgeorge> Eighth_Doctor: yes but we don't need to give the detractors more ammo
20:35:25 <carlwgeorge> nirik: we don't promise anything for epel either
20:35:44 <Eighth_Doctor> we don't promise anything period
20:35:54 <carlwgeorge> perceptions matter, whether they are based on fact or speculation
20:35:55 <tdawson> How about if we approve 7 days for each, and I get a script written (I just found part of it) that will give us some stats over the past year or two.
20:35:58 <nirik> indeed.
20:36:10 <Eighth_Doctor> fine
20:36:21 * nirik is fine with 7/7 or 7/3 either one...
20:36:41 <Eighth_Doctor> it's just a pain when windows are tight to get rebuilds done between point releases
20:36:54 <tdawson> I'm fine with either, but I'm quite curious about the stats.
20:36:59 <Eighth_Doctor> and Bodhi chokes on the KDE side tags
20:37:14 <pgreco> I'd say let's start with 7/7 and do another check in a few weeks
20:37:23 <Eighth_Doctor> which means we have to lengthen the time to rebase/update things quite a lot
20:37:49 <Eighth_Doctor> I think tdawson did ~5 separate Bodhi updates to bring up Plasma in epel-next
20:38:05 <Eighth_Doctor> which is mostly because doing it all in one go makes Bodhi cry
20:38:06 <nirik> pgreco: sounds reasonable
20:38:12 <Eighth_Doctor> and nobody is doing anything about Bodhi
20:38:16 <tdawson> Eighth_Doctor: Oh, that was because in the past, bodhi would choke on anything over 100 packages
20:38:38 <Eighth_Doctor> it still doesn't like it very much :)
20:39:02 <tdawson> No, no it doesn't, and I got a couple timeout errors with 150, but it did it. :)
20:39:33 <Eighth_Doctor> but basically, large desktop stack updates are painful in EPEL because of those problems
20:39:41 <Eighth_Doctor> and shortening the window makes it less painful
20:39:42 <tdawson> Anyway, before we go over time ... All in favor of 7 days for all, for now, and another vote in 2 weeks on if we change -next to something lower.
20:40:08 <Eighth_Doctor> +1
20:40:13 <tdawson> +1
20:40:19 <pgreco> +1
20:40:35 <nirik> +1
20:40:47 <nirik> https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-NEXT-2021-2409835866 has 146. :)
20:41:46 <pgreco> nirik, for the next recheck, is there a way you can get metrics on usage of epel vs epel-testing?
20:41:50 <tdawson> #info Change bodhi update to 7 days for all of epel, with the provision that we discuss and vote again about lowering epel-next wait time, in two weeks:  5 (yes) 0 (no)
20:42:50 <tdawson> OK, one more voting, and or opinion thing, which hopefully will be quicker
20:43:05 <nirik> pgreco: thats very hard to say... I think there might be some stats for it in the stuff mattdm_ collects... not sure
20:43:17 <tdawson> EPEL Logo preliminary vote - https://pagure.io/design/issue/770#comment-744088
20:44:04 <nirik> what exactly are we voting on there?
20:44:11 <tdawson> Of the 6 logos in that list, which do people like the most
20:44:42 <tdawson> This isn't the end logo, just to figure out what direction to take the logo.
20:44:58 <carlwgeorge> sorry was pulled away, i'm +1 on the 7/7 days
20:45:27 <pgreco> tdawson: off by one error, it starts at 0 ;)
20:45:32 <pgreco> so there are 7 options
20:45:38 <tdawson> Ha!
20:45:40 <nirik> I think 5 and 6
20:45:54 <nirik> or rather
20:46:00 <pgreco> I like (6) mostly because of the reference to the E
20:46:13 <pgreco> as it is depicted in the next message
20:46:28 <Eighth_Doctor> I like (3) and (4)
20:46:56 <Eighth_Doctor> I lean slightly toward (3) because it makes it look more symbiotic, but I'm also fine with (4) for the reasons carlwgeorge likes it
20:47:05 <carlwgeorge> out of those options i like 4 the most
20:47:22 <carlwgeorge> > "Rotate (4)" makes me think "RHEL base, extended by Fedora", which is nice.
20:47:40 <tdawson> nirik you never finished your sentance "or rather ..."
20:48:04 <nirik> I meant: 5 OR 6, not 5 and 6 since I think you want us to vote on one. ;)
20:48:26 <nirik> I think I like 5 best as it's more simple /clear than 6...
20:48:37 <tdawson> I was actually thinking of asking people to vote on 2, but give which of those two they prefer
20:49:18 <pgreco> ok, I'll go with 6 first, 4 second
20:49:30 <nirik> 5 first, 4 second
20:49:59 <tdawson> Eighth_Doctor: Were you 3 first, 4 second ?
20:50:05 <Eighth_Doctor> yees
20:50:07 <Eighth_Doctor> *yes
20:50:17 <pgreco> I'm seeing a pattern here...
20:50:29 <nirik> ha
20:50:57 <tdawson> carlwgeorge: I know your 4 first, any prefernce as for your second vote?
20:51:14 <carlwgeorge> 2 but not rotated
20:52:00 <tdawson> And I'm 3 first, 4 second, but really fine with either.
20:52:32 <tdawson> It's looking like 4 is the one that people either like, or are ok with.
20:52:47 <pgreco> yeap
20:52:47 <nirik> it's growing on me as I look at it more.
20:53:25 <tdawson> I just saw the time.   And I know that pgreco has one thing for open floor.
20:53:46 <pgreco> it's just a comment, so no worries
20:54:16 <tdawson> OK, anything for EPEL 7 or EPEL 8 ?
20:55:06 <tdawson> The one thing I have a concern with is Eighth_Doctor's bug about the groups getting their defaults stripped ...
20:55:26 <tdawson> That didn't come out right ... let me get the bug
20:55:30 <Eighth_Doctor> https://pagure.io/releng/issue/10228
20:55:38 <nirik> yeah, not sure where that bug is...
20:56:14 <tdawson> Original bug was here https://bugzilla.redhat.com/show_bug.cgi?id=1984706
20:57:23 <tdawson> Basically the kde-desktop-enviroment isn't pulling in what it's supposed to pull in.
20:58:03 <tdawson> Anyway, I'll add more to the issue so it's easier to see what's going on, I just haven't had a chance to yet.
20:58:50 <nirik> we talked about it in the last releng meeting, jednorozec was going to investigate and see if it was a pungi bug or what.
20:59:06 <tdawson> ok thanks.
20:59:26 <tdawson> #topic General Issues / Open Floor
20:59:42 <tdawson> pgreco: OK, your on. :)
21:00:03 <pgreco> ok, so I talked to maha, and from what I gather, the idea is to start the updates on rawhide
21:00:27 <pgreco> so the incompatible updates and notifications will start there, and then move its way down to stable fedoras and epel
21:00:43 <pgreco> so we'll know more once the dates are closer, but everything seems to be on track
21:00:57 <pgreco> that's it
21:01:24 <tdawson> pgreco: and maha knows to send emails to epel-devel?  and not just fedora-devel ... err  devel@
21:01:50 <pgreco> yeap, but I'll get in touch again after rawhide
21:01:54 * nirik isn't following, which updates?
21:01:58 <tdawson> OK, thank you
21:02:15 <tdawson> Oh, this is a discussion from last week, it's about tor
21:02:31 <Eighth_Doctor> lol fedora-devel
21:02:33 <pgreco> nirik tor is completely removing its already deprecated torv2 addressing system
21:02:34 <nirik> ah, ok. I wasn't here last week. ;)
21:02:44 * Eighth_Doctor remembers when devel@ actually was fedora-devel@
21:03:06 <tdawson> Ha!
21:03:19 <pgreco> and that's going to generate some issues on people, so we want the maintainer to send some notifications in order to try to lower those
21:03:27 <pgreco> but nothing we can do to stop it
21:03:51 <nirik> fair enough
21:03:58 <tdawson> Our time is us.  Thank you everyone for the good discussions, voting, and helping make EPEL great.
21:04:23 <tdawson> We'll talk next week here, if not somewhere else sooner. :)
21:04:32 <pgreco> see ya!, bye
21:04:32 <tdawson> #endmeeting