15:00:00 <bowlofeggs> #startmeeting Bodhi stakeholders (2017-09-12)
15:00:00 <zodbot> Meeting started Tue Sep 12 15:00:00 2017 UTC.  The chair is bowlofeggs. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:00 <zodbot> The meeting name has been set to 'bodhi_stakeholders_(2017-09-12)'
15:00:00 <bowlofeggs> #meetingname bodhi_stakeholders
15:00:00 <zodbot> The meeting name has been set to 'bodhi_stakeholders'
15:00:00 <bowlofeggs> #topic salutations
15:00:00 <bowlofeggs> #chair bowlofeggs caleigh dgilmore masta mboddu nirik pbrobinson puiterwijk Kellin jcline
15:00:00 <zodbot> Current chairs: Kellin bowlofeggs caleigh dgilmore jcline masta mboddu nirik pbrobinson puiterwijk
15:00:19 * mboddu is kinda here
15:00:44 * jcline is here
15:00:59 * nirik is kinda here too.
15:02:04 <dustymabe> .hello dustymabe
15:02:06 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
15:03:14 <bowlofeggs> thanks for coming everyone. let's do this
15:03:22 <bowlofeggs> #topic announcements and information
15:03:22 <bowlofeggs> #info A Bodhi 2.11.0 beta is deployed to staging, but may become 2.12.0, because:
15:03:22 <bowlofeggs> #info threebean hopes to release and deploy a 2.11.0 that does module mashing, displacing the current planned 2.11.0
15:03:22 <bowlofeggs> #info Release notes available at https://bodhi.stg.fedoraproject.org/docs/release_notes.html
15:03:39 <bowlofeggs> so those release notes are for what may become 2.12.0...
15:04:22 <bowlofeggs> any comments/questions on the announcements?
15:04:38 <Southern_Gentlem> explain your info posts above please
15:05:04 <bowlofeggs> Southern_Gentlem: what in particular is not clear?
15:05:16 <Southern_Gentlem> so is he deploying 2.11.0 beta
15:05:55 <bowlofeggs> Southern_Gentlem: what had been planned to be a 2.11.0 was deployed already, but threebean wants to "take" that version number and make it a release that adds nothing but module mashing
15:05:56 * threebean waves
15:06:06 <bowlofeggs> Southern_Gentlem: if that happens, what was going to be 2.11 will instead become 2.12
15:06:17 <threebean> well - i'm glad to take any version number that makes sense.  bowlofeggs suggested that I take 2.11.0.
15:06:23 <bowlofeggs> basically, i didn't want to introduce both changes in an FBR deployment
15:06:34 <Southern_Gentlem> 2.11.1?
15:06:38 <bowlofeggs> yeah we don't *have* to use that version, but 2.11 wasn't taken yet
15:06:52 <bowlofeggs> Southern_Gentlem: bodhi uses semantic versioning, so a .z release should not introduce a feature
15:07:03 <dustymabe> ehh 2.12 works for me
15:07:05 <dustymabe> :)
15:07:06 <bowlofeggs> and 2.11.0 isn't actually taken yet (there's no tag for it)
15:07:14 <dustymabe> ahh ok, then w/e
15:07:40 <Southern_Gentlem> ok its clear as mud
15:07:46 <Southern_Gentlem> continue
15:07:47 <bowlofeggs> we also aren't completely sure this will happen, as we need an FBR approved to do the deployment
15:08:08 <bowlofeggs> but if threebean's testing goes well and the FBR is approved, that's the plan
15:08:24 <bowlofeggs> so actually, the only topic i had planned for today was that
15:08:25 <bowlofeggs> so:
15:08:28 <threebean> :p
15:08:30 <bowlofeggs> #topic module mashing
15:09:07 <bowlofeggs> normally i like to avoid deploying features during beta freezes, but this particular feature is an important fedora server objective
15:09:23 <bowlofeggs> so that is why we are doing things that might be considered "strange" :)
15:11:00 <bowlofeggs> any other thoughts or questions about bodhi and modularity?
15:11:06 <dustymabe> bowlofeggs: yeah
15:11:12 <dustymabe> so the PR for this in bodhi
15:11:36 <dustymabe> seems to 'hardcode' a lot of the config that would have traditionally been in the pungi config file
15:11:40 * threebean nods
15:11:45 <dustymabe> like
15:11:47 <dustymabe> https://github.com/fedora-infra/bodhi/pull/1697/files#diff-1a7f8244263b14fccf760eb233aedd61R1168
15:11:58 <dustymabe> I don't really like including information like that in "code"
15:12:12 <dustymabe> or is that just in the tests?
15:12:14 <threebean> yeah - mcurlej is pulling that out.  we thought we had another week or so to do it, but the fesco request has put some urgency on it.
15:12:28 <dustymabe> threebean: pulling that out means?
15:12:29 <bowlofeggs> dustymabe: that appears to be a test file
15:12:36 <threebean> replacing it with configurable values.
15:12:47 <dustymabe> threebean: how would it be configured?
15:13:09 <dustymabe> bowlofeggs: right that is a test file, but how would those values get defined otherwise?
15:13:09 <threebean> dustymabe: well - the line you linked to?  that's just in the tests.
15:13:27 <threebean> "bodhi-manage-releases create"
15:13:46 <threebean> they're database values.
15:13:57 <dustymabe> oh wow, so the pungi config lives in the database?
15:14:02 <threebean> no.
15:14:09 <threebean> the "Release" object lives in the database.
15:14:14 * threebean checks the link again
15:14:16 <bowlofeggs> dustymabe: the bit you linked to was a Release object
15:15:01 <dustymabe> ok, i guess what I'm trying to figure out is how this work would merge with the 'call pungi from bodhi' work that I have worked on
15:15:01 <bowlofeggs> which is a database object - this isn't new for modularity but exists today as well (it's how Fedora 26 is defined today, for example)
15:15:14 <dustymabe> and if they overlap at all
15:15:26 <bowlofeggs> dustymabe: yeah i also share that concern, because they currently are very different
15:15:53 <dustymabe> i was just trying to verify that we weren't storing the pungi config by hard coding
15:16:02 <dustymabe> if that's not the case then 'good'
15:16:11 <threebean> that is 100% not the plan.  :)
15:16:25 <dustymabe> threebean: where would the pungi configs be stored?
15:16:39 <threebean> likely in /etc/bodhi/
15:16:52 <threebean> managed by ansible, like the mash configs
15:17:05 <threebean> second option - we could store them in a pagure repo, like the comps files.
15:17:08 <dustymabe> threebean: ok let's collaborate on that
15:17:20 <threebean> +1
15:17:38 <dustymabe> threebean: https://pagure.io/releng/issue/6971
15:17:41 <dustymabe> please comment
15:17:56 <threebean> dustymabe: we're flat out this week just trying to get the crazy release branch rebased and to verify upgrade/downgrade works.
15:18:07 <bowlofeggs> threebean: it's important to me not to spread bodhi's configs into more places - i.e., we should keep them in the ansible repo like we do with the mash configs
15:18:22 <threebean> +1
15:18:33 <threebean> no prob with that from me.
15:19:23 <dustymabe> threebean: so can we collaborate on what the name of the files would be
15:19:32 <dustymabe> to make sure we don't step on each others toes
15:20:35 <nirik> wasn't there a releng ticket about the config here? or was it something else.
15:20:37 * nirik looks
15:20:50 <dustymabe> nirik: https://pagure.io/releng/issue/6971
15:21:09 <nirik> yeah. that one
15:21:37 <threebean> dustymabe: yeah, for sure.  I'll make sure to put comments in that ticket.
15:22:53 <bowlofeggs> threebean: care to give an update on the status of the PR?
15:24:25 <threebean> no, I haven't gotten into it yet.  I've been spending all my time setting up staging to test traditional bodhi.
15:25:05 <threebean> I was planning to try and rebase "whatever" is in the PR right now ontop of 2.10.0, cut a 2.11.0.beta3 (or something like that) and use that to verify that upgrade/downgrade of bodhi doesn't break the traditional bodhi mash.
15:25:06 <dustymabe> threebean: including mash?
15:25:12 <threebean> dustymabe: yes.
15:25:45 <threebean> I successfully mashed a repo of "Fedora 2000 updates-testing" in staging.  :)
15:25:47 <dustymabe> bowlofeggs: ^^
15:25:55 <dustymabe> i didn't think that was possible
15:26:02 <threebean> dustymabe: it took a lot of work.
15:26:19 <dustymabe> threebean: sigh :(
15:26:26 <bowlofeggs> threebean: ah ok. i've been working on https://github.com/fedora-infra/bodhi/issues/1793 and i'm getting close to done - there's a few bits that i need to use your fancy pants infer_content_class() that are slightly tricky, but i should finish it today
15:26:40 <threebean> bowlofeggs: rad :)
15:26:41 <dustymabe> i wish that 'lot of work' could have been targeted towards the bodhi-calling-pungi work
15:26:43 <bowlofeggs> threebean: congrats on fedora 2000 :)
15:27:25 <bowlofeggs> threebean: were you able to verify if the resulting repo can be used to update a module?
15:27:35 <threebean> no, because the resulting repo was for traditional bodhi.
15:27:41 <bowlofeggs> ahhhh
15:27:47 <bowlofeggs> oh, i thought that had a module in it?
15:27:58 <threebean> no, i was only trying to replicate how bodhi works today.
15:28:00 <bowlofeggs> threebean: i guess you wanted to verify that traditional works before and after?
15:28:01 <threebean> so that I can test it.
15:28:02 <bowlofeggs> yeah ok
15:28:04 <bowlofeggs> i thinki understand
15:28:05 <bowlofeggs> cool
15:28:07 <bowlofeggs> makes sense
15:28:07 <threebean> so that I can then upgrade it, and verify that it *still* works.
15:28:12 <bowlofeggs> indeed
15:28:15 <threebean> and then downgrade it, and verify that it *still* works.
15:28:22 <threebean> and then upgrade it, and verify that *modules* work.
15:28:29 <bowlofeggs> cool, i think that will help persuade the FBR gods
15:29:49 <bowlofeggs> any other thoughts on this topic, or shall we move to open floor?
15:30:52 <bowlofeggs> #topic open floor
15:31:51 <bowlofeggs> <crickets>
15:31:53 <threebean> <3
15:32:06 <bowlofeggs> going once!
15:32:10 <bowlofeggs> going twice!
15:32:19 <bowlofeggs> sold, to threebean for an FBR!
15:32:22 <bowlofeggs> #endmeeting