#meeting-1:fedoraproject.org log

<@tdawson:fedora.im>

18:00:09

!startmeeting EPEL (2025-02-12)

<@meetbot:fedora.im>

18:00:10

Meeting started at 2025-02-12 18:00:09 UTC

<@meetbot:fedora.im>

18:00:10

The Meeting name is 'EPEL (2025-02-12)'

<@tdawson:fedora.im>

18:00:15

!topic aloha

<@tdawson:fedora.im>

18:00:15

!meetingname epel

<@meetbot:fedora.im>

18:00:16

The Meeting Name is now epel

<@carlwgeorge:fedora.im>

18:00:18

!hi

<@zodbot:fedora.im>

18:00:20

Carl George (carlwgeorge) - he / him / his

<@tdawson:fedora.im>

18:00:26

Hi Carl George

<@dherrera:fedora.im>

18:00:33

!hi

<@zodbot:fedora.im>

18:00:35

Diego Herrera (dherrera) - he / him / his

<@tdawson:fedora.im>

18:00:44

Hi Diego Herrera

<@jonathanspw:fedora.im>

18:01:48

!hi

<@zodbot:fedora.im>

18:01:50

Jonathan Wright (jonathanspw)

<@tdawson:fedora.im>

18:02:16

Hi Jonathan Wright

<@jonathanspw:fedora.im>

18:02:21

howdy howdy

<@tdawson:fedora.im>

18:03:32

We're a little light on people today ... maybe more will come.

<@tdawson:fedora.im>

18:05:29

!topic EPEL Issues https://pagure.io/epel/issues

<@tdawson:fedora.im>

18:05:29

!link https://pagure.io/epel/issues?tags=meeting&status=Open

<@salimma:fedora.im>

18:05:52

!hi

<@zodbot:fedora.im>

18:05:53

Michel Lind (salimma) - he / him / his

<@tdawson:fedora.im>

18:06:04

Hi Michel Lind 🎩 UTC-6

<@salimma:fedora.im>

18:06:06

today's full of meetings left and right

<@tdawson:fedora.im>

18:06:29

The only open issue we have marked with meeting is the DNF variable one. Do we want to go over that right now, or wait for the EPEL10 section of the meeting?

<@tdawson:fedora.im>

18:06:38

Ah, let's go over it now.

<@jonathanspw:fedora.im>

18:06:40

every wednesday for me :(

<@conan_kudo:matrix.org>

18:06:45

!hi

<@zodbot:fedora.im>

18:06:48

Neal Gompa (ngompa) - he / him / his

<@tdawson:fedora.im>

18:06:51

!313

<@davide:cavalca.name>

18:06:54

!hi

<@tdawson:fedora.im>

18:06:57

!epel 313

<@carlwgeorge:fedora.im>

18:06:57

yeah i can give a status update

<@zodbot:fedora.im>

18:06:57

Davide Cavalca (dcavalca) - he / him / his

<@zodbot:fedora.im>

18:06:58

● **Opened:** 3 weeks ago by carlwgeorge

<@zodbot:fedora.im>

18:06:58

<@zodbot:fedora.im>

18:06:58

**epel #313** (https://pagure.io/epel/issue/313):**EPEL 10 dnf variables problems**

<@zodbot:fedora.im>

18:06:58

● **Last Updated:** 4 days ago

<@zodbot:fedora.im>

18:06:58

● **Assignee:** carlwgeorge

<@conan_kudo:matrix.org>

18:07:10

my Wednesdays suck too

<@tdawson:fedora.im>

18:07:20

Hi Davide Cavalca and Conan Kudo

<@carlwgeorge:fedora.im>

18:08:21

all the dnf/libdnf prs have been merged upstream. backports to the el10 packages have been proposed, and we're working through the process to get those merged and included. this will be for both problems described in the issue.

<@conan_kudo:matrix.org>

18:09:19

sweetness

<@salimma:fedora.im>

18:09:24

nice

<@carlwgeorge:fedora.im>

18:09:51

oh and the necessary provides have been added to redhat-release as well, so that part is done

<@conan_kudo:matrix.org>

18:10:01

woot

<@tdawson:fedora.im>

18:10:07

This fixes the packagekit problems too?

<@conan_kudo:matrix.org>

18:10:08

so we just wait now

<@carlwgeorge:fedora.im>

18:10:47

yes, the plan is to include that fix at the same time so they're qa tested together

<@tdawson:fedora.im>

18:10:57

Very nice

<@tdawson:fedora.im>

18:11:10

Thank you, and anyone else, for pushing those through.

<@carlwgeorge:fedora.im>

18:12:07

it's still possible for the backports to be rejected, so we're not completely out of the woods, but so far it's trending in the right direction

<@tdawson:fedora.im>

18:12:35

Understood.

<@salimma:fedora.im>

18:12:52

we just proposed the centos fast track SIG - so if it got rejected we can always temporarily park it there

<@carlwgeorge:fedora.im>

18:13:58

one possibility is it's just declined for 10.0 and deferred to 10.1, in which case it would be good test case for the fasttrack sig

<@tdawson:fedora.im>

18:14:08

The people being affected are RHEL users ... I'm not sure having it in a CentOS SIG is going to really help. But it's a thought.

<@salimma:fedora.im>

18:14:30

that is a good point

<@salimma:fedora.im>

18:14:35

well, it depends on how we set up the SIG

<@salimma:fedora.im>

18:14:52

if there is a need we can ask Fabian to give us the ability to build against RHEL

<@tdawson:fedora.im>

18:14:53

What I mean, is that centos users, it's simple to install the fastrack sig repo, but for RHEL users ... not as easy.

<@carlwgeorge:fedora.im>

18:15:20

centos users are affected by the packagekit problem

<@salimma:fedora.im>

18:15:25

e.g. have the ability to target minor releases like epel10.Y and not just targeting centos stream + epel10

<@salimma:fedora.im>

18:15:37

but yeah it will make the onboarding for EPEL even more annoying

<@conan_kudo:matrix.org>

18:16:01

let's cross that bridge if we have to come to it

<@tdawson:fedora.im>

18:16:06

Oh, I didn't realize packagekit was a problem for everyone.

<@conan_kudo:matrix.org>

18:16:10

and assume for now everything will land

<@conan_kudo:matrix.org>

18:16:29

the fact everything is done and being processed is a great sign

<@carlwgeorge:fedora.im>

18:16:37

yeah the releasever_minor stuff is only a problem for rhel, but the packagekit libdnf thing affects everyone

<@tdawson:fedora.im>

18:17:11

OK, I'll remain positive :) ... and do what I can to get things moving along.

<@tdawson:fedora.im>

18:18:11

Anyway, that is great news that things are moving in the right direction. Anything else before we move on?

<@tdawson:fedora.im>

18:19:05

I'll take that as a no, and move on.

<@tdawson:fedora.im>

18:19:14

!topic EPEL 10

<@tdawson:fedora.im>

18:19:43

The one thing I have for this is the minor branching SOP

<@tdawson:fedora.im>

18:19:58

!epel 304

<@zodbot:fedora.im>

18:19:59

● **Assignee:** dherrera

<@zodbot:fedora.im>

18:19:59

**epel #304** (https://pagure.io/epel/issue/304):**EPEL 10.0 mass branching**

<@zodbot:fedora.im>

18:19:59

<@zodbot:fedora.im>

18:19:59

● **Opened:** 3 months ago by carlwgeorge

<@zodbot:fedora.im>

18:19:59

● **Last Updated:** 20 minutes ago

<@tdawson:fedora.im>

18:20:31

Diego Herrera: Do you have an update?

<@dherrera:fedora.im>

18:21:30

yeah :) kevin gave lot's of feedback on the SOP Draft yesterday, and we are working with Carl George in testing those on staging

<@carlwgeorge:fedora.im>

18:22:28

our goal is to do the mass branching in staging this week, and prod next week

<@carlwgeorge:fedora.im>

18:22:43

we've been getting lots of great feedback so far from the releng folks

<@dherrera:fedora.im>

18:23:12

I think that the only things that I have questions about are some ansible playbooks I need to check and see if we need to block access to infra (which might not be necesary, but we'll see), but besides that, it depends on how things behave on staging

<@tdawson:fedora.im>

18:24:27

I know the branching is the hardest part, and thank you all for all the work on that, but I have another tagging question.

<@tdawson:fedora.im>

18:24:55

When we do the tagging, will we be bringing over the packages in -testing as well as those that are released?

<@carlwgeorge:fedora.im>

18:25:08

that actually came up in the releng channel yesterday

<@dherrera:fedora.im>

18:25:15

I think that the only things that I have questions about are some ansible playbooks I need to check, and see if we need to block access to infra during the process (which might not be necesary since it's shorter than the fedora one, but we'll see), but besides that, it depends on how things behave on staging

<@tdawson:fedora.im>

18:26:01

My mind keeps going back and forth on it.

<@carlwgeorge:fedora.im>

18:26:05

my initial thought was that yes all those should be tagged over as well to ensure a proper upgrade path, but there may be some with negative karma

<@nhanlon:beeper.com>

18:26:21

!hi

<@zodbot:fedora.im>

18:26:23

Neil Hanlon (neil) - he / him / his

<@nhanlon:beeper.com>

18:26:31

late, sry.. double booked myself

<@tdawson:fedora.im>

18:26:46

Well, ya, those with negative karma, like k3b, .... hmm

<@tdawson:fedora.im>

18:26:54

Hi Neil Hanlon

<@carlwgeorge:fedora.im>

18:27:23

right now there are 98 updates in bodhi for epel 10.0 in pending or testing status

<@carlwgeorge:fedora.im>

18:27:47

fedora avoids this problem because rawhide doesn't have a testing period

<@salimma:fedora.im>

18:28:39

the maintainer can always push it to stable after a week right?

<@tdawson:fedora.im>

18:28:58

But if they push it to stable, then they've missed the tagging.

<@carlwgeorge:fedora.im>

18:29:02

yes, but that would only go into epel10.0, and would miss epel10.1

<@carlwgeorge:fedora.im>

18:29:22

i'm thinking it should be possible to write a script that uses the bodhi api to check for pending/testing updates for epel10.0, and tag the relevant koji builds to epel10.1 as long as there is no negative karma

<@carlwgeorge:fedora.im>

18:29:42

it wouldn't be foolproof, but it would probably catch most of what we want to happen

<@salimma:fedora.im>

18:29:46

ahh.. yeah I see the problem

<@tdawson:fedora.im>

18:29:59

Besides k3b ... how many others have negative karma? Or is that hard to tell without going through each one?

<@tdawson:fedora.im>

18:30:11

Because I could just drop k3b if that's the only one.

<@carlwgeorge:fedora.im>

18:30:38

i know of a half dozen or so

<@dherrera:fedora.im>

18:30:51

we could automatically drop the ones that have negative karma as part of the process...

<@carlwgeorge:fedora.im>

18:31:02

interesting idea...

<@carlwgeorge:fedora.im>

18:31:52

like unpush the ones in testing with negative karma at the time of branching, and if the maintainers submit a new build it will be for epel10.1

<@carlwgeorge:fedora.im>

18:32:44

that would also be a good periodic cleanup for updates left dangling without maintainer action

<@tdawson:fedora.im>

18:32:50

Yep, I've been thinking of doing that for k3b.

<@carlwgeorge:fedora.im>

18:33:24

now, what do we do with updates that have no karma, work fine, but just haven't been in testing for a full week? fast track them to stable so they get tagged for 10.1?

<@salimma:fedora.im>

18:33:52

I think tag them into 10.1 since it's more experimental while letting them do the normal thing for 10.0?

<@salimma:fedora.im>

18:34:15

on the ground that 10.1 is more experimental and we care more about availability

<@carlwgeorge:fedora.im>

18:34:24

yeah maybe, we could manually tag those for 10.1 but let 10.0 happen normally

<@salimma:fedora.im>

18:34:28

or... if you really want to be sure, auto create a new update for 10.1 that is also in testing

<@salimma:fedora.im>

18:34:37

match the current state. idk how hard that would be to automate

<@conan_kudo:matrix.org>

18:35:16

I would probably just let it land in 10.0 and tag up to 10.1 later

<@tdawson:fedora.im>

18:36:51

I guess one question is if it's possible to script something that skips the negative karma packages.

<@carlwgeorge:fedora.im>

18:37:17

in theory, sure

<@tdawson:fedora.im>

18:37:49

If it is, then I think we should skip tagging them up. I'm not adamant about that, but it's my opinion right now.

<@smooge:fedora.im>

18:38:02

morning

<@tdawson:fedora.im>

18:38:11

Morning Stephen J Smoogen

<@salimma:fedora.im>

18:38:22

morning

<@carlwgeorge:fedora.im>

18:38:30

we're also overdue for a policy that says maintainers are responsible for ensuring a proper upgrade path between minor versions. that came up at connect in an audience question. that factors into this as well.

<@jonathanspw:fedora.im>

18:39:13

Shouldn't it be a non-issue since maintainers shouldn't be breaking things within a major EL version?

<@carlwgeorge:fedora.im>

18:39:42

it's not about breaking changes, it's about the upgrade path in general

<@jonathanspw:fedora.im>

18:40:11

I think the question at connect was about upgrading between major versions, not minors. I could be mis-remembering

<@tdawson:fedora.im>

18:40:18

So, if we don't tag things in -testing correctly, it's possible that 10.0 has a higher number package than what's in 10.1

<@salimma:fedora.im>

18:40:23

yeah. if the package is not tagged in 10.1 but is in 10.0 that breaks the upgrade path

<@salimma:fedora.im>

18:40:58

will 10.1 allow updates to go without karma at the beginning? or will we turn on gating from the get go

<@carlwgeorge:fedora.im>

18:41:02

the way it came up at connect was what if a maintainer did a change in the epel10.0 branch, but didn't also build it in the epel10 branch. that could result in say foo-1.0-5.el10_0 that is higher than foo-1.0-4.el10_1

<@jonathanspw:fedora.im>

18:41:15

Oh, I'm definitely mis-remembering. ignore me.

<@jonathanspw:fedora.im>

18:41:19

Was conflating it with a different question

<@carlwgeorge:fedora.im>

18:41:24

same thing would happen in fedora if a maintainer applied a change to f41 but didn't do it in rawhide/f42

<@salimma:fedora.im>

18:41:25

right. we should probably just adopt Fedora's policy here

<@carlwgeorge:fedora.im>

18:41:51

yeah modeled after fedora's policy for sure, probably with some wording tweaks

<@salimma:fedora.im>

18:41:54

we see that even in Fedora sometimes - maintainers might want to push updates to stable before working out how to deal with F42/Rawhide FTBFSes

<@salimma:fedora.im>

18:42:17

even with a policy in place. so without a policy I could imagine it being the wild west. no bueno for an EL repo :)

<@tdawson:fedora.im>

18:42:57

I think this needs an issue plus email and/or discourse

<@carlwgeorge:fedora.im>

18:43:09

i can start that thread today

<@jonathanspw:fedora.im>

18:43:14

I have to step out, y'all have a good rest of the week!

<@tdawson:fedora.im>

18:43:33

Thank you Carl George

<@tdawson:fedora.im>

18:43:58

Looking at the time, I think I need to timebox this topic and move it to the issue.

<@tdawson:fedora.im>

18:44:19

Now that Smooge is here, I feel it's time to move on to the next topic

<@tdawson:fedora.im>

18:44:27

!topic Old Business

<@salimma:fedora.im>

18:45:19

I see smooge immediately has something to say

<@smooge:fedora.im>

18:45:35

and forgets everyone can see him typing

<@tdawson:fedora.im>

18:45:52

Any other old business?

<@smooge:fedora.im>

18:46:03

not from me.

<@carlwgeorge:fedora.im>

18:46:18

i see an old issue hanging around

<@carlwgeorge:fedora.im>

18:46:23

!epel 303

<@zodbot:fedora.im>

18:46:24

● **Opened:** 4 months ago by fmaurer

<@zodbot:fedora.im>

18:46:24

<@zodbot:fedora.im>

18:46:24

● **Assignee:** Not Assigned

<@zodbot:fedora.im>

18:46:24

● **Last Updated:** 3 months ago

<@zodbot:fedora.im>

18:46:24

**epel #303** (https://pagure.io/epel/issue/303):**Can stgit be updated in epel9**

<@tdawson:fedora.im>

18:47:29

It seems like they were going to do more investigation, and then replied back.

<@tdawson:fedora.im>

18:47:38

Or have you heard or seen anything else from them?

<@carlwgeorge:fedora.im>

18:47:42

looks like it's still an issue in git

<@carlwgeorge:fedora.im>

18:47:51

i haven't seen anything outside of this ticket

<@tdawson:fedora.im>

18:48:08

It seems like they were going to do more investigation, and then haven't replied back.

<@carlwgeorge:fedora.im>

18:48:15

nothing on the list either

<@tdawson:fedora.im>

18:48:33

Maybe we can give them a ping, see what they decided.

<@carlwgeorge:fedora.im>

18:48:39

yeah i'll update the ticket

<@tdawson:fedora.im>

18:48:44

Thank you.

<@tdawson:fedora.im>

18:48:50

Anything else?

<@tdawson:fedora.im>

18:49:46

!topic General Issues / Open Floor

<@tdawson:fedora.im>

18:49:57

Does anyone have anything for Open Floor?

<@tdawson:fedora.im>

18:51:44

If not, I'm ok ending the meeting a little early.

<@salimma:fedora.im>

18:51:58

I'm ok having lunch a bit less late :)

<@tdawson:fedora.im>

18:52:18

Then let's call this a meeting. :)

<@tdawson:fedora.im>

18:52:41

Thank you everyone for the good discussions. And thank you very much for your work on EPEL and all you do for it's community.

<@tdawson:fedora.im>

18:52:52

I'll talk to you next week, if not sooner.

<@zodbot:fedora.im>

18:53:09

neil has already given cookies to tdawson during the F41 timeframe

<@tdawson:fedora.im>

18:53:30

!endmeeting