17:06:11 #startmeeting RELENG (2018-02-01) 17:06:11 Meeting started Thu Feb 1 17:06:11 2018 UTC. The chair is Kellin. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:06:11 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:06:11 The meeting name has been set to 'releng_(2018-02-01)' 17:06:27 morning. 17:06:28 #meetingname releng 17:06:28 The meeting name has been set to 'releng' 17:06:32 morning 17:06:50 * Kellin is using his logs to find commands, so bear with me as I learn on the fly how to drive :) 17:07:07 #chair dgilmore nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin 17:07:07 Current chairs: Kellin dgilmore masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll 17:09:07 * Kellin checks for issues tagged meeting 17:10:49 #topic "Broken dependencies" report does not understand rich dependencies 17:11:12 #link https://pagure.io/releng/issue/6365 17:12:13 well, yes this is a thing. Someone needs to re-write it. ;) (not me) 17:13:34 do we know if Pungi developed the ability to replace it? 17:14:12 not that I know of. 17:14:45 what is spam-o-matic's intended use case again? 17:15:03 it lets developers know that their package(s) have broken deps in rawhide and need to be fixed. 17:16:18 there is a dnf repoclosure plugin... might be possible to leverage that 17:17:26 possibly yeah - though it works completely different to what we had before. I'll tag it for taiga and get it worked into our todo list, it may already be there in a different form 17:17:49 yeah, might be worth pinging pungi folks too... in case they did implement something along these lines. 17:19:10 updated ticket 17:19:56 #info Need to reach out to Pungi developers to see if this got added to their codebase; need to add this into Taiga to include in our work planning. 17:21:11 #topic [RFE] Add new repositories for modular[-updates[-testing]] 17:21:12 #link https://pagure.io/releng/issue/7227 17:22:14 I think this is waiting on the pungi work finishing. 17:23:20 it looks like the final word was to use a script to merge and skip pungi changes though 17:24:18 huh, ok. who is going to write that tho? 17:24:56 * nirik thought pungi folks already were working on adding this. 17:25:00 it doesn't specify, and I'm not savvy enough on what it would entail to /merge/ a compose to know if that's even safe 17:25:27 https://pagure.io/pungi/pull-request/830 17:26:09 sounds like lsledar is going to test it and then get it merged. 17:26:56 #info Looks like the pungi updates have been submitted in a PR: https://pagure.io/pungi/pull-request/830 17:28:26 #topic Initial discussion for rpm-ostree jigdo 17:28:32 #link https://pagure.io/releng/issue/7201 17:29:19 it seems we need to book an actual meeting for this 17:29:26 there's a discussion chain here, but not a lot of movemnet 17:29:33 *movement 17:30:06 yeah, dunno here... 17:33:09 #info Updated ticket tagging involved parties to see if we can get a kickoff meeting scheduled to move this forward 17:33:28 #topic Update OSBS for Flatpak Support 17:33:36 #link https://pagure.io/releng/issue/6998 17:34:44 I guess cverna is on the hook for this now? but not sure the status 17:35:21 so I'm a little confused 17:35:26 we have a 'release-engineering' github? 17:35:30 https://github.com/release-engineering 17:37:08 hum... 17:37:16 that might be the factory2.0 folks? 17:37:29 or I guess Red Hat releng? 17:38:14 factory has its own repo set in pagure I thought 17:38:42 nirik: Kellin there is still some work for that 17:39:15 first step is to have the latest upstream packaged in Fedora 17:39:53 which I am currently working on, then we need to deploy OSBS in orchestrated mode which seems quite challenging 17:40:51 * dustymabe slides in 17:41:12 cverna: so the million dollar question - do you think there's any hope of landing in F28 or is this a 'down the road' item? 17:41:20 cverna: is that the many arches? or something else... 17:42:16 Kellin: I would say that there is a good chance for it too land in F28 I think this is otaylor whish too 17:43:20 nirik: yes the OSBS architecture for multi arches is needed for Flatpak 17:43:31 yeah, tricky. ;( 17:44:23 We have roughly 4 months - that seems like a lot to get done, validate, test and deploy. (note that's release final F28 in four months) 17:44:44 if it's multi-arch 17:45:16 I am not super familiar with this architecture yet, The only thing is that internal (Red Hat) OSBS team has this setup deployed 17:45:49 although they did not the test Faltpak support, but they have multi-arch 17:46:01 so I'll mark it as potentially yellow for Proj. Mgmnt since you're still working out the scopes/details - is that fair? 17:46:12 sounds good to me 17:48:24 #info Upstream pull requests are merged. @cverna actively working on pulling these updates into Fedora from upstream. Scoping on the work still happening due to complexities in delivering multi-arch. Current status for F28 is yellow. 17:48:53 +1 17:48:54 #topic automatically put the latest composed Modular Fedora containers into the Fedora candidate registery 17:49:01 #link https://pagure.io/releng/issue/7147 17:49:25 okay - anyone know what 'lgtm' means? :) 17:49:52 looks good to me 17:49:53 looks good to me 17:50:11 nirik++ 17:50:13 * nirik is reminded he needs a f27 mirrorlist container 17:51:08 so pushing to the candidate registry - is that like rawhide for containers? 17:51:25 nirik: does that mean the current mirrorlist machine needs to be rebuild on f27 ? 17:52:26 cverna: no, just need to build and use a newer container. ;) 17:52:37 Kellin: more like updates-testing I would think 17:53:15 do we have an SOP or process defined for what goes into that registry already? 17:53:28 is this ticket needed anymore? we didn't end up making a f27 modular server. 17:54:11 I thought we moved to a different hybrid process? 17:54:34 stripped down Fedora core operating system with modules on top 17:54:54 yes, that was the pungi ticket eariler. 17:54:54 at least for F28 17:54:57 it's not done yet 17:55:10 and this looks like old f27 stuff... I think this ticket can just be closed. 17:55:46 I believe OSBS will push to candidate registry once we have bodhi able to do container update 17:56:02 so we think this can be closed for now? 17:56:16 cverna: do you concur with nirik ? 17:56:28 then bodhi will move the images from candidate to production registry 17:57:10 * cverna reads the ticket again :) 17:57:48 * nirik nods 17:58:02 so conclusion is - F27 Modular did not make it to production delivery. Mode of delivery changed for F28. OSBS will push to the registery once bodhi can do container updates obsoleting the request in this ticket. ? 17:58:04 I would say that for this to happen it still requires some work (change OSBS to push to candidate registry) 17:58:17 and have bodhi to manage container update 17:58:59 Kellin: agreed with your conclusion 17:59:34 OK - posting that up 17:59:41 #info F27 Modular did not make it to production delivery. Mode of delivery changed for F28. OSBS will push to the registry once bodhi can do container updates obsoleting the request in this ticket. 18:01:11 should I make it closed as "cant fix" or "invalid" or other? 18:01:15 not sure what status 18:01:40 invalid I guess 18:01:48 nirik: cverna dustymabe any pressing issues? 18:01:53 #topic Open Floor 18:02:08 not on my side 18:02:19 Kellin: nothing pressing :) 18:02:26 always fires to put out though :) 18:02:31 mass rebuild is coming up. 18:02:45 and we need a new fedora-release package 18:02:48 otherwise nothing 18:03:01 okay, nirik let's discuss that in #releng since it's already time here - but good call 18:03:16 #info nirik and kellin to discuss mass-rebuild and fedora-release package 18:03:19 Thanks everyone 18:03:22 #endmeeting