17:05:40 <smooge> #startmeeting EPEL (2018-11-21)
17:05:40 <zodbot> Meeting started Wed Nov 21 17:05:40 2018 UTC.
17:05:40 <zodbot> This meeting is logged and archived in a public location.
17:05:40 <zodbot> The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:05:40 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:05:40 <zodbot> The meeting name has been set to 'epel_(2018-11-21)'
17:06:00 <smooge> #meetingname epel
17:06:00 <zodbot> The meeting name has been set to 'epel'
17:06:01 <smooge> #topic Chair and Introductions
17:06:01 <smooge> #chair avij bstinson Evolution nirik smooge
17:06:01 <zodbot> Current chairs: Evolution avij bstinson nirik smooge
17:06:09 <nirik> morning
17:06:19 <smooge> #topic Open Floor
17:06:42 <pgreco> hello
17:06:43 <smooge> I did not send out an agenda so this meeting is open floor.
17:07:32 <pgreco> ideas for epel8?
17:07:43 <pgreco> to continue with the ML threads....
17:09:00 * nirik would like to make a lot modular...
17:09:29 * pgreco would like to understand how this modular thing will work
17:10:29 <avij> hello
17:10:48 <nirik> well, like fedora has it already. ;)
17:10:52 <pgreco> all I have in mind, is that we have too many branches for packages, they'll deviate too much from fedora
17:11:01 <pgreco> and will be difficult to keep up
17:11:36 <avij> smooge: last meeting you asked if we could keep the one-hour-later timeslot, but now we're again at the regular timeslot ..
17:11:38 <pgreco> so I think it would be easier to keep all in a single branch, with macros
17:11:39 * avij is confused
17:11:42 <avij> but here
17:12:11 <smooge> avij, I am here because I didn't change the meeting schedule in the invite or the topic in #epel
17:12:23 <nirik> pgreco: right, thats what stream branches do... but not clear if we can build them against rhel8 easily.
17:12:25 <smooge> so I figured I would confuse people if I did a different time
17:12:40 <smooge> I confuzed people either way
17:12:41 <avij> alright. let's go on with the show then ..
17:12:57 <smooge> so first off. is 1800 UTC ok for future meetings?
17:13:03 <smooge> I will change the two right now
17:13:03 <avij> yes
17:13:06 <smooge> done
17:14:37 <nirik> sure.
17:15:17 <bstinson> works for me, i kind of forgot about this timeslot myself
17:15:47 <pgreco> I'm in front of the computer all day :), ok either way
17:15:50 <smooge> ok I have fixed the calendar and the topic
17:16:22 <smooge> OK so we started syncing the beta content this week
17:17:00 <smooge> I would like to make the beta channels for EPEL-8.0-beta to be base, appstream, codeready and not have high availability in the builds
17:17:40 * nirik hasn't looked, but ok
17:17:44 <smooge> I would also like to make two EPEL-8.0-beta subchannels: base and appstream
17:17:59 <smooge> base is where non modular content goes. appstream is where modules go
17:18:30 <smooge> I will write this more up onto the mailing lists
17:19:40 <nirik> all sounds reasonable.
17:19:48 <smooge> anything else for today? Next week we will meet at 1800 UTC and that will be our permanent time
17:20:04 <nirik> it would be nice to have a easy guide for maintainers as to which they should try and use...
17:20:10 <avij> I'm worried if this modularity requires more package maintainer manpower. but there's little we can do about that.
17:20:40 <nirik> and without ursa-major 'base' packages may not be too many.
17:20:42 <avij> yeah, a quickstart guide to modularity for EPEL package maintainers would be "nice"
17:20:45 <smooge> avij, oh it does
17:21:11 <smooge> I spent the last 2 days trying to read through the docs and I am as lost as I was when I started
17:21:31 <smooge> as far as I can tell they are not really meant for 'singular' packagers
17:21:43 <smooge> they are meant to be teams of people
17:22:30 <nirik> that may be true, but we can also leverage the fedora ones... provided things build they don't have to do much to enable a epel8 build of their existing modle
17:22:31 <nirik> module
17:22:37 <smooge> I mean a singular packager can try and do them but it takes more work than the old 'just touch my package next week'
17:24:36 <pgreco> well, how much of this can be forced on packagers?
17:24:51 <smooge> packagers are like cats
17:25:11 <pgreco> I mean, checking that the master builds on epel automagically
17:25:19 <pgreco> like many packagers do
17:25:41 <smooge> pgreco, I don't think it will be possible without a '8-rawhide'
17:25:55 <nirik> you mean us test building all the modules? or ?
17:26:40 <smooge> but nirik's question is probably better than me answering
17:27:04 <pgreco> nirik, I'm not exactly sure how, but trying to get epel as close as possible with fedora
17:27:27 <pgreco> for me it is always easier to add a few macros, than to keep a separate branch completely
17:27:33 <nirik> I'm not sure... there is a module-build-local so we could perhaps use that to test...
17:28:27 <nirik> perhaps we could invite some modularity knowledageable folks next week?
17:28:50 <avij> sounds like a good idea
17:29:14 <avij> especially if they have an idea for how to fit EPEL into this modularity equation
17:30:37 <nirik> It would be really good to know if fedora modules will just build ok in epel8... or if they will need completely new branches or tweaks
17:30:44 <smooge> who would that be?
17:30:53 <bstinson> the modularity question gets stickier too because, yes, the versions are all available in parallel, but the migration of streams in and out of RHEL is going to be interesting
17:30:54 <nirik> possibly sgallagh or contyk ?
17:31:04 <contyk> yes?
17:31:15 * sgallagh looks up
17:31:17 <nirik> contyk: hey. Talking epel8 and modules. :)
17:31:25 <sgallagh> We should have some of those
17:31:38 <nirik> yeah. I think so!
17:31:45 <smooge> contyk, sgallagh would you be available next week at 1800 to talk about how this would be implemented and what packagers do
17:31:56 <contyk> yes
17:31:59 <sgallagh> 1800 UTC?
17:32:03 <smooge> UTC
17:32:16 <nirik> what are the chances that fedora modules would just build on epel8? or would we need -epel8 stream branches and different module-md?
17:32:18 <bstinson> we might need to do this in a few sessions...because we (CentOS) are also just starting to try to figure out what building downstream modules might look like
17:32:26 <smooge> we are needing to have people who are doing the same things they have done since EL5 to know what is going on with EL8
17:32:36 <smooge> we will definitely need multiple sessions
17:33:12 <contyk> the chances are pretty high
17:33:19 <contyk> at least these days; might change in ten years
17:33:31 <kwizart> rpm-build seems to belong to a module...
17:33:40 <nirik> I think the number of packages we can have in epel8 as base packages will be pretty small really (until we have ursa major anyhow)
17:34:12 <nirik> ha. then we might need it to have any non modular packages.
17:34:39 <bstinson> now is also the time to discuss breaking some of our 10-year "obligations" as well...
17:34:50 <smooge> well I thought we broke those years ago
17:34:57 <bstinson> well formally at least
17:35:07 <smooge> ok
17:35:24 <nirik> +1
17:35:25 <smooge> yeah we need to formally declare what we are doing and how we are going to do it
17:35:29 <nirik> sla's all around.
17:35:33 <smooge> sle's
17:35:38 <bstinson> slo's?
17:35:47 <smooge> agreements require signatures
17:35:52 <nirik> elo's?
17:35:54 <bstinson> what's the other one?
17:36:00 <nirik> expectations
17:36:04 <smooge> service level expectations
17:36:17 <smooge> that is where we say what you can expect from us
17:36:21 <bstinson> expectations, objectives, and there's an i-word
17:36:22 * nirik was just using what pdc called it
17:36:48 <nirik> ie, "I will maintain this for 1 year"
17:37:49 <smooge> avij, bstinson nirik I have to head out sooner than I realized
17:38:11 <smooge> aka I have 2 people sitting tapping their feet outside my office
17:38:11 <nirik> lets take things to list/next week...
17:38:25 <smooge> thanks
17:38:27 <bstinson> +1 i'll bring questions
17:38:41 <smooge> actually can you send some of them to the list so I can have an agenda
17:38:44 <smooge> thanks
17:39:23 <smooge> if there isn't anything esle.. I am going to call this meeting over now
17:39:38 <smooge> thank you all for coming.. and my apologies for ducking out
17:39:42 <smooge> #endmeeting