17:08:15 <mboddu> #startmeeting RELENG (2017-09-14)
17:08:15 <zodbot> Meeting started Thu Sep 14 17:08:15 2017 UTC.  The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:08:15 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:08:15 <zodbot> The meeting name has been set to 'releng_(2017-09-14)'
17:08:15 <mboddu> #meetingname releng
17:08:16 <zodbot> The meeting name has been set to 'releng'
17:08:16 <mboddu> #chair dgilmore nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin
17:08:16 <zodbot> Current chairs: Kellin dgilmore masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll
17:08:16 <mboddu> #topic init process
17:08:45 <Kellin> .hello kellin
17:08:46 <zodbot> Kellin: kellin 'None' <kellin@retromud.org>
17:09:22 <tyll> .hello till
17:09:22 <dgilmore> hi
17:09:23 <zodbot> tyll: till 'Till Maas' <opensource@till.name>
17:09:27 <nirik> morning
17:09:37 <tyll> hi
17:09:48 <puiterwijk> hello
17:11:44 <mboddu> Okay, first of all sorry about the channel change. I will look into the other options so that we wont clash with other meetings.
17:11:57 <mboddu> Lets get started
17:12:31 <mboddu> nirik: IIRC, last time you said you have couple of things to talk, is it still valid, then I would like to start with Open Floor
17:12:33 <mboddu> ?
17:13:27 <nirik> I... don't recall. ;)
17:13:44 <mboddu> nirik: okay, then lets get started with meeting topics
17:14:51 <mboddu> #topic #6998 Update OSBS for Flatpak Support
17:14:58 <mboddu> #link https://pagure.io/releng/issue/6998
17:15:12 <dustymabe> .hello dustymabe
17:15:13 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
17:15:22 <dustymabe> o/
17:15:36 <mboddu> I want to bring this up so that we can discuss on what needs to be done on RelEng/Infra side so that we can start supporting flatpaks and deliver them
17:16:19 <mboddu> I dont want it have any hiccups like we had with modularity
17:19:04 <dustymabe> nobody with any thoughts?
17:19:15 <dustymabe> seems like adam would have some
17:19:17 * nirik hasn't been following this. who has been working on it?
17:19:28 <mboddu> dustymabe: seems like everyone is busy with Go/No-Go
17:20:18 <mboddu> nirik: I guess Factory 2.0
17:20:36 <maxamillion> .hello maxamillion
17:20:37 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
17:20:53 <maxamillion> sorry I'm late, is this a permanent time slot change?
17:21:12 <dustymabe> mboddu: mind recounting the current topic?
17:21:26 <nirik> maxamillion: no, just overlap with go/nogo
17:21:59 <dustymabe> nirik: i thought the time change is permanaent
17:22:05 <dustymabe> i.e. this is the time this meeting will run
17:22:10 <mboddu> maxamillion: its a permanent change time wise but we got a conflict with Go/No-Go in #fedora-meeting-2 for today
17:22:37 <nirik> yes, but not channel. I suppose it could be channel too since irc support sig hasn't been meeting lately.
17:23:29 <maxamillion> mboddu: ok, so permanent time change but not channel
17:23:34 * maxamillion updates his calendar entry
17:23:40 <mboddu> maxamillion: yes
17:23:59 <maxamillion> alright, so ... Flatpak in OSBS
17:26:39 <maxamillion> ?
17:27:12 <dustymabe> maxamillion: right that is the topic
17:27:23 <dustymabe> 13:15:36           mboddu | I want to bring this up so that we can discuss on what needs to be done on RelEng/Infra side so that we can start supporting flatpaks and deliver them
17:27:29 <dustymabe> 13:16:19           mboddu | I dont want it have any hiccups like we had with modularity
17:27:44 <mboddu> maxamillion: yes
17:28:17 <dgilmore> afaik flatpak code is still being actively worked out
17:28:24 <mboddu> maxamillion: seems like its a lot of work on your side - koji container build, osbs client and atomic reactor are the pieces needs updating?
17:28:33 <dgilmore> it relies on modules being built and shipped
17:28:51 <maxamillion> oh man
17:28:52 <maxamillion> so
17:29:40 <mboddu> dgilmore: AFAIK flatpak work will be done in a month(there is 1 PR and couple of bug fixes to fix). They are guessing they will get it done by next month
17:29:42 <maxamillion> basically otaylor had done all the work in various upstreams to make it work, the build and release stuff is going to look almost identical to how we do containers ... the big question in my mind is bodhi, but that's still a question for containers as well
17:29:48 <maxamillion> has done*
17:30:23 <mboddu> maxamillion: yes, bodhi needs updating as well
17:30:45 <maxamillion> there's still an outstanding patch for the docker-distribution registry to handle OCI images and some metadata format extensions (iirc), but the Fedora maintainer of the package agreed to carry those patches and update as it makes it's way into mainline, so I'm not concerned there
17:30:54 <maxamillion> other patches in OSBS have been merged or are in review
17:30:55 <mboddu> dgilmore, maxamillion : Another thing I am not sure of is, how we deliver them?
17:31:06 <maxamillion> mboddu: registry.fedoraproject.org
17:31:10 <maxamillion> mboddu: just like the container images
17:31:47 <dustymabe> maxamillion: +1
17:31:53 * otaylor_ is around, but things seem to be understood well - nothing to add at the moment
17:32:39 <maxamillion> also, otaylor++ for all that work
17:32:39 <zodbot> maxamillion: Karma for otaylor changed to 1 (for the f26 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:33:00 <dgilmore> mboddu: the container registry afaik
17:33:02 <maxamillion> this is the first fedora feature ever in OSBS that I didn't have to do anything for :)
17:33:24 <mboddu> maxamillion: haha :)
17:33:33 <otaylor_> Making this smooth in the end for updates will eventually require freshmaker support, auto-preparing updates of containers/flatpaks, notifying maintainers
17:33:59 <mboddu> otaylor_: When do you think all the work upstream will be done, by next month?
17:34:04 <maxamillion> otaylor_: +1
17:34:29 <mboddu> otaylor_: And thanks for all of your help
17:35:03 <maxamillion> mboddu: so the upstream work will likely be done by then, but it's a matter of deployment into stage OSBS ... my priorities have been shuffled a bit because we're blocked on a couple things for OSBS currently and there's going to be a time period where I'm waiting on a new team member to join who will be taking over FLIBS
17:35:03 <nirik> otaylor++ Thanks for all your work on this
17:35:03 <zodbot> nirik: Karma for otaylor changed to 2 (for the f26 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:35:34 <otaylor_> mboddu: I think everything is at the state where it *could* be deployed in Fedora infrastructure today - though there's still more improvements to do now that the bulk of the changes landed in OSBS this morning
17:35:43 <otaylor_> (that was blocking smaller changes)
17:36:32 <otaylor_> The main thing that's not ready now is an index for registry.fedoraproject.org to enable installation from GNOME Software - I have some code written for that, but it needs considerable consultation with different stakeholders to figure out exactly what it should look like
17:37:07 <otaylor_> that's not going to block the rest, just until that is ready trying out built flatpaks will require some hoop-jumping
17:37:38 <mboddu> otaylor_, maxamillion : thanks for the info
17:38:16 <mboddu> Seems like its going to take a while to get into fedora space
17:39:00 <mboddu> maxamillion: So, bodhi, koji container build, osbs client and atomic reactor needs updating. Am I missing anything?
17:39:09 <maxamillion> otaylor_: oh yeah, the indexing thing
17:39:22 <otaylor_> mboddu: probably depends the cycles that maxamillion will have to help with the deployment side of deployment and thus how painful multi-arch keeps on being :-)
17:40:15 <maxamillion> otaylor_: multi-arch is blocked, I'm waiting until after the next sprint to continue, the stuff I was expecting to land in the latest release isn't ready yet, so I'm just putting it on hold until that stuff lands
17:42:37 <mboddu> otaylor_, maxamillion : seems like a lot of work at our hands
17:43:00 <mboddu> Anyway, thanks for the info guys.
17:43:16 <maxamillion> certainly
17:43:18 <maxamillion> and yes, it is :/
17:43:22 <maxamillion> lots of work
17:44:10 <otaylor_> mboddu: my goal is to keep this to be as upstream and simple-to-deploy as possible, but nothing that involves feeding things through koji and openshift with a dose of modularity thrown in is ever *simple*
17:44:46 <mboddu> otaylor_: yep :D
17:45:05 <maxamillion> otaylor_: +1
17:45:15 <mboddu> #info Upstream work will be done with in a month or so. But in Fedora land, we need to update lots of tools like bodhi, koji container build, osbs client and atomic reactor to support it.
17:46:59 <maxamillion> +1
17:48:18 <mboddu> #topic Alternative Architectures updates
17:48:27 <mboddu> sharkcz: if you are around? Any updates?
17:51:46 <mboddu> I gues sharkcz is not around
17:51:55 <mboddu> So, open floor time
17:52:01 <mboddu> #topic Open Floor
17:52:08 <mboddu> Anybody has anything to share?
17:52:14 <nirik> oh, I did have one thing...
17:52:39 <mboddu> nirik: hehe, you remembered ;)
17:52:40 <nirik> we had storage slowdowns a few weeks ago... then we got them to go away, but were seeing massive i/o...
17:53:07 <nirik> we finally tracked that down to the rawhide-composer vm... I rebooted it the other day and the weird high i/o went away.
17:53:17 <nirik> so, hopefully we are all back to normal now.
17:54:05 <mboddu> nirik: Do you think of anything that might have caused it?
17:54:36 <nirik> well, not sure. I tried unmounting and remounting things and it didn't matter... but some nfs interaction between f26 and the netapp I guess.
17:55:00 <nirik> The slowdowns we are pretty sure are caused by nfs 4.1 and netapp... we moved all our mounts back to 4.0
17:55:25 <nirik> (to be clear, the bug might be in linux kernel/nfs client side or netapp (server side) we don't know yet)
17:56:28 <mboddu> #info The issue with massive i/o has been tracked down to rawhide-composer vm and Kevin rebooted it and it fixed the issue. The bug might be in linux kernel/nfs client side or netapp server side.
17:56:31 <mboddu> nirik: thanks
17:57:07 <mboddu> Anybody got anything else?
17:57:21 <maxamillion> nirik: that sounds like a pain to diagnose
17:57:40 <nirik> yeah, folks are working on it.
17:59:39 <mboddu> Okay, that is it guys, thanks for joining
17:59:43 <mboddu> #endmeeting