serversig
LOGS
21:00:14 <sgallagh> #startmeeting Fedora Server SIG Weekly Meeting (2017-12-19)
21:00:14 <zodbot> Meeting started Tue Dec 19 21:00:14 2017 UTC.  The chair is sgallagh. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:14 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
21:00:14 <zodbot> The meeting name has been set to 'fedora_server_sig_weekly_meeting_(2017-12-19)'
21:00:14 <sgallagh> #meetingname serversig
21:00:14 <zodbot> The meeting name has been set to 'serversig'
21:00:14 <sgallagh> #topic Roll Call
21:00:18 <sgallagh> .hello2
21:00:19 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
21:00:28 <smooge> .hello2
21:00:29 <zodbot> smooge: smooge 'Stephen J Smoogen' <smooge@gmail.com>
21:00:42 <smooge> Thinks this will be a lonely meeting
21:00:49 * langdon sorta here
21:00:52 <dperpeet> .hello dperpeet
21:00:53 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
21:01:49 <sgallagh> dperpeet: What, really? Didn't your wife just have a child? Why are you here?
21:02:21 * smooge goes for the 1950's joke.. "His wife just had a child"
21:02:30 <dperpeet> sgallagh, it's night here, no child of mine likes to sleep at night when they're young :)
21:02:37 <sgallagh> Well, we don't have a lot to discuss anyway.
21:02:42 <sgallagh> #topic Agenda
21:03:08 <sgallagh> #info Agenda Item: Hybrid Modularity
21:03:17 <sgallagh> anyone else want to bring up a topic?
21:03:30 <dperpeet> * Server testing
21:03:42 <dperpeet> hmm
21:03:53 <dperpeet> should I rephrase with info?
21:03:59 <sgallagh> #char dperpeet
21:04:02 <sgallagh> #chair dperpeet
21:04:02 <zodbot> Current chairs: dperpeet sgallagh
21:04:08 <sgallagh> (sorry, didn't mean to cook you)
21:04:15 <dperpeet> #info Agenda Item: Fedora Server testing
21:04:23 <dperpeet> no problem, cuts down the heating costs
21:04:36 <sgallagh> ha!
21:04:44 <sgallagh> OK, let's get rolling
21:04:57 <sgallagh> #topic Hybrid Modularity
21:04:59 <smooge> I read your article
21:05:12 <sgallagh> #link https://communityblog.fedoraproject.org/modularity-dead-long-live-modularity/
21:06:39 <sgallagh> tl;dr: We've come up with a way to adapt the modularity plan such that it can live atop the traditional Fedora release rather than fully replacing it.
21:06:46 <smooge> yay
21:06:48 <sgallagh> Though replacing it may remain a long-term goal
21:07:36 <sgallagh> I submitted this blog post as a first introduction to the new plan. I'll be writing a series of more technical posts after the New Year to describe the actual implementation in greater detail.
21:07:52 <dperpeet> this will have an interesting impact on CI, probably
21:07:59 <sgallagh> But I wanted to get this information out to the general public before the holidays so that people have time to internalize this informatino.
21:08:44 <smooge> I expect we will need to bring it up a couple of times
21:09:43 <sgallagh> I've also sent a text-only version to the devel-announce list just now.
21:10:15 <smooge> mainly because people have internalized multiple different versions of how modularity is going to do X (or fail to)
21:10:18 <sgallagh> There are several big advantages to this new approach, not least of which is that it can be applied to all of Fedora at once. So Server Edition doesn't need to go it alone any more
21:10:24 <sgallagh> yes
21:11:01 <sgallagh> Anyway, the only action item here is
21:11:31 <sgallagh> #info Server SIG members should read this blog entry and come prepared to ask questions at the Jan. 2 meeting
21:11:50 <sgallagh> #topic Server Testing
21:11:58 <sgallagh> dperpeet: You have the floor
21:12:08 <dperpeet> thanks, I will try not to leave scorch marks
21:12:12 <dperpeet> (being charred and all)
21:12:34 <dperpeet> I'm still working on CI
21:12:49 <dperpeet> and my question is, do we want to have any CI specific to Fedora Server as a whole?
21:12:57 <dperpeet> a lot of tests are focused on specific packages
21:13:08 <dperpeet> but there might be a few overarching, important stories
21:13:13 <dperpeet> that we can link to changes in multiple repos
21:13:18 <dperpeet> err... packages
21:13:41 <dperpeet> so if there are stories / tests like that, I would like to know about them
21:14:05 <dperpeet> with the goal that they are run on all dist-git changes of involved packages
21:14:39 <sgallagh> dperpeet: So, that's a really good question, but I think it might be best started as a thread on the list.
21:14:48 <dperpeet> sure
21:14:51 <sgallagh> And then in a future meeting we can winnow it down to the set we want to focus on.
21:14:55 <dperpeet> I don't think this has to be discussed here
21:15:02 <dperpeet> but I wanted to put it out as food for thought
21:15:28 <dperpeet> you can also file things like that under https://pagure.io/fedora-ci/AtomicCi/
21:15:32 <dperpeet> don't let the atomic fool you :)
21:16:02 * sgallagh nods
21:16:28 <sgallagh> I guess the question I have is how this would work. Would this be post-compose testing?
21:16:39 <sgallagh> And if so, what is different from openqa?
21:16:46 <dperpeet> the CI pipeline will update the compose for every change
21:16:58 <dperpeet> and we want to move this towards pull request testing
21:17:11 <sgallagh> ok
21:17:19 <sgallagh> Define a "change" in this context, please
21:17:29 <dperpeet> anything that changes dist-git
21:17:37 <dperpeet> for a (set of) packages
21:17:57 <dperpeet> e.g. rebasing a package, such as sed, to a new version
21:18:01 <dperpeet> or adding a patch
21:18:12 <dperpeet> or changing the spec file
21:18:13 <dperpeet> et cetera
21:19:11 <sgallagh> I can certainly see this being a boon to dealing with the FreeIPA stack, if nothing else.
21:19:18 <sgallagh> (Since it touches so many pieces)
21:19:37 <dperpeet> good point
21:19:42 <dperpeet> I'll keep that in mind as one story
21:19:43 <sgallagh> Ditto Cockpit, but I kind of assume dperpeet has that covered already
21:20:11 <dperpeet> I hope I won't forget Cockpit, heh
21:20:32 <dperpeet> the suggestion of the mailing list is good, I'll follow that in the new year
21:23:06 <dperpeet> that's it from me on that topic :)
21:23:09 <sgallagh> dperpeet: If possible, start the thread beforehand. Give people something to think about while avoiding their least-favorite relatives :)
21:23:53 <dperpeet> heh
21:23:58 <dperpeet> I will try, but no promises
21:25:08 <sgallagh> #topic Open Floor
21:25:08 <smooge> one of the items I would like to work out before hand is how the package flow works in the CI. The various CI/koschei/etc tools have really picked up the amount of diskspace used for logging
21:25:14 <sgallagh> Whoops.
21:25:19 <smooge> no rpoblem
21:25:23 <sgallagh> #topic Server Testing
21:25:40 <smooge> this is a long term request versus a "oh we need to solve this in this meeting"
21:25:52 <sgallagh> #chair smooge
21:25:52 <zodbot> Current chairs: dperpeet sgallagh smooge
21:25:53 <dperpeet> yes, but data retention is indeed an important topic
21:26:00 <sgallagh> smooge: Mind capturing the request in the minutes?
21:26:12 <dperpeet> I have a few thoughts on this
21:26:24 <dperpeet> and some experience from Cockpit testing
21:26:38 <smooge> #info Need to work out network webflow of how CI works so that logging is done in a way which is useful for metrics
21:26:54 <smooge> #info also so that we can budget disk space needs and such
21:28:08 <smooge> that was it
21:28:11 <dperpeet> :)
21:28:17 <sgallagh> OK, anything else on this topic?
21:28:24 <dperpeet> in short, existing tools shouldn't be affected much currently
21:28:38 <smooge> thank you for rewinding over that
21:28:39 <dperpeet> but this deserves more attention in the future
21:30:14 <sgallagh> OK
21:30:20 <sgallagh> #topic Open Floor
21:30:44 <sgallagh> #info There will be no meeting on 2017-12-26
21:30:57 <sgallagh> Other Open Floor topics?
21:31:10 <smooge> Happy New Year?
21:31:36 <smooge> For those on a European calendar
21:32:20 <dperpeet> yeah, and a Merry Christmas :)
21:32:22 <sgallagh> And to anyone celebrating a holiday over the next couple weeks, enjoy yourselves!
21:32:46 <smooge> Be kind to one another, for we all have to sober up on Monday
21:33:56 <smooge> I think we are done here :)
21:35:21 <sgallagh> Thanks for coming, folks!
21:35:23 <sgallagh> #endmeeting