<@jbtrystram:matrix.org>
16:31:43
!startmeeting fedora_coreos_meeting
<@meetbot:fedora.im>
16:31:48
Meeting started at 2024-09-18 16:31:43 UTC
<@meetbot:fedora.im>
16:31:48
The Meeting name is 'fedora_coreos_meeting'
<@jbtrystram:matrix.org>
16:31:53
!topic roll call
<@jbtrystram:matrix.org>
16:31:57
!hi
<@zodbot:fedora.im>
16:31:59
Jean-Baptiste Trystram (jbtrystram) - he / him / his
<@aaradhak:matrix.org>
16:32:52
!hi aaradhak
<@zodbot:fedora.im>
16:32:53
Aashish Radhakrishnan (aaradhak)
<@jaimelm:fedora.im>
16:33:06
!hi
<@zodbot:fedora.im>
16:33:08
Jaime Magiera (jaimelm)
<@hricky:fedora.im>
16:33:12
!hi
<@zodbot:fedora.im>
16:33:14
Hristo Marinov (hricky) - he / him / his
<@siosm:matrix.org>
16:33:50
!hi
<@zodbot:fedora.im>
16:33:53
Timothée Ravier (siosm) - he / him / his
<@dustymabe:matrix.org>
16:34:21
!hi
<@zodbot:fedora.im>
16:34:23
Dusty Mabe (dustymabe) - he / him / his
<@jbtrystram:matrix.org>
16:35:21
alright let's get started
<@jbtrystram:matrix.org>
16:35:37
!topic Action items from last meeting
<@jbtrystram:matrix.org>
16:36:03
!info marmijo to rebase `next-devel` to F41 and stop the `branched` stream
<@marmijo:fedora.im>
16:36:23
!hi
<@zodbot:fedora.im>
16:36:26
Michael Armijo (marmijo)
<@marmijo:fedora.im>
16:36:44
This was done on Friday 👍️
<@jbtrystram:matrix.org>
16:37:09
nice ! Thanks marmijo
<@jbtrystram:matrix.org>
16:37:13
marmijo: ++
<@zodbot:fedora.im>
16:37:19
jbtrystram gave a cookie to marmijo. They now have 4 cookies, 3 of which were obtained in the Fedora 40 release cycle
<@dustymabe:matrix.org>
16:37:33
marmijo:++
<@zodbot:fedora.im>
16:37:36
dustymabe has already given cookies to marmijo during the F40 timeframe
<@marmijo:fedora.im>
16:37:43
:)
<@zodbot:fedora.im>
16:38:00
dustymabe gave a cookie to zodbot. They now have 27 cookies, 7 of which were obtained in the Fedora 40 release cycle
<@jbtrystram:matrix.org>
16:38:11
!info the FCOS `next-devel` stream rebased on fedora 41 content on september 13th
<@dustymabe:matrix.org>
16:38:35
oh nice - I didn't know adding the cookie emoji did the same thing as `++`
<@zodbot:fedora.im>
16:38:46
siosm gave a cookie to zodbot. They now have 28 cookies, 8 of which were obtained in the Fedora 40 release cycle
<@apiaseck:matrix.org>
16:38:47
!hi
<@zodbot:fedora.im>
16:38:50
Adam Piasecki (c4rt0) - he / him / his
<@zodbot:fedora.im>
16:38:51
dustymabe has already given cookies to marmijo during the F40 timeframe
<@zodbot:fedora.im>
16:39:09
siosm gave a cookie to marmijo. They now have 5 cookies, 4 of which were obtained in the Fedora 40 release cycle
<@siosm:matrix.org>
16:39:26
so many 🍪!
<@jbtrystram:matrix.org>
16:39:29
!info gursewak and marmijo to organize a test day for F41 for the week after 41 beta is released with help from Renata Ravanelli
<@zodbot:fedora.im>
16:39:49
jbtrystram gave a cookie to siosm. They now have 28 cookies, 6 of which were obtained in the Fedora 40 release cycle
<@dustymabe:matrix.org>
16:39:51
welp F41 beta was released yesterday :)
<@marmijo:fedora.im>
16:40:13
yes, that was faster than we expected
<@jbtrystram:matrix.org>
16:40:30
nothing prevents starting to organize I guess :D
<@jbtrystram:matrix.org>
16:42:50
let's put that on next week actions I guess
<@jbtrystram:matrix.org>
16:43:16
marmijo: gursewak do you need more time ? We can push back the test day a week maybe ?
<@marmijo:fedora.im>
16:44:31
gursewak: and I can get together to organize it, but we haven't begun that yet. I'm not sure how much time is needed.
<@dustymabe:matrix.org>
16:45:15
Some time is needed just to announce so people can plan to join
<@dustymabe:matrix.org>
16:45:37
First step would be contacting [@sumantrom:fedora.im](https://matrix.to/#/@sumantrom:fedora.im)
<@gurssing:matrix.org>
16:47:23
+1: we can definitely get started with it then.
<@jbtrystram:matrix.org>
16:47:31
I'll reschedule the action for next week then and we can move on to the topics ?
<@jbtrystram:matrix.org>
16:47:54
!action gursewak and marmijo to organize a test day for F41 for the week after 41 beta is released with help from Renata Ravanelli
<@jbtrystram:matrix.org>
16:48:19
We have one topic today :
<@jbtrystram:matrix.org>
16:48:20
!topic Collaboration with OKD Working Group to Build SCOS Artifacts
<@jbtrystram:matrix.org>
16:48:34
<@jaimelm:fedora.im>
16:49:33
Quick Context for folks not familiar: OKD is transitioning to a more community-supported model. Part of that is taking ownership of the SCOS image builds in addition to the cluster artifact builds.
<@jaimelm:fedora.im>
16:50:06
Things are quite scattered and we would like to find a more permanent home for building SCOS.
<@jaimelm:fedora.im>
16:50:30
In our minds, it makes sense to reach out to the CoreOS build pros :)
<@jaimelm:fedora.im>
16:50:45
The rest you can glean from the issue.
<@siosm:matrix.org>
16:52:29
There are 3 parts to this:
<@siosm:matrix.org>
16:52:29
- infra to run the pipeline on (needs an Kubernetes/OCP cluster right now)
<@siosm:matrix.org>
16:52:29
- cloud credentials to push AMIs, etc.
<@siosm:matrix.org>
16:52:29
- an S3 bucket to push things to
<@siosm:matrix.org>
16:52:59
We could re-use some of those from what we have in FCOS, but should we?
<@jaimelm:fedora.im>
16:53:40
We currently have OKD builds happening on MOC, which is OpenShift and Tekton. We're an outlier though compared to the other CoreOS variants.
<@jaimelm:fedora.im>
16:54:05
We're not sure how long the MOC stuff will be tenable.
<@dustymabe:matrix.org>
16:54:21
is the plan for SCOS to standalone, or just be a implementation detail of OKD?
<@jaimelm:fedora.im>
16:55:01
Initially, it was to be a detail of OKD. However, the community has been asking for standalone.
<@jaimelm:fedora.im>
16:55:38
The people who decided it would be a detail have all moved on from OKD. So, that decision was inherited initially.
<@siosm:matrix.org>
16:55:49
For me, if this is standalone, it really has to be happening as part of the CentOS project (not necessarily the infra, but at least the project part)
<@dustymabe:matrix.org>
16:56:30
yeah, it being standalone (I'd love to have it exist) is a lot of work and requires a lot of extra thought/effort
<@dustymabe:matrix.org>
16:56:52
for example, do you have streams? what is your release cadence? etc..
<@jaimelm:fedora.im>
16:57:23
We're working that out now.
<@dustymabe:matrix.org>
16:57:43
in OKD, or part of a new community initiative ?
<@jaimelm:fedora.im>
16:58:48
The decision to switch to SCOS was somewhat sudden. All OKD FCOS builds were halted. So, we're working on the first OKD SCOS release now. After the cluster release, we'll have more time to shift focus to the SCOS aspect.
<@jaimelm:fedora.im>
16:59:48
The initiative would be for now part of the OKD community, but part of the impetus for doing stand alone is that we'd get more people contributing to the project as a whole if they could contribute to the SCOS part.
<@jaimelm:fedora.im>
17:01:26
This is part of positioning OKD as a true upstream of OCP (as much as possible/allowed)
<@dustymabe:matrix.org>
17:02:28
<@dustymabe:matrix.org>
17:02:28
I think not knowing for sure the end state makes it a bit harder to know the right path forward for now.
<@dustymabe:matrix.org>
17:02:39
but we never really know the future, do we? :)
<@dustymabe:matrix.org>
17:03:05
travier: Jonathan Lebon : thoughts?
<@jaimelm:fedora.im>
17:03:16
Right. I think our concern is that we're not sure how long MOC will be an option and ultimately we want to position ourselves in such a way that we're not adrift in separate infrastructure, culture, etc.
<@dustymabe:matrix.org>
17:03:33
travier: Jonathan Lebon anyone: thoughts?
<@dustymabe:matrix.org>
17:03:42
travier: Jonathan Lebon: anyone: thoughts?
<@dustymabe:matrix.org>
17:04:06
is MOC serviceable for now? i.e. are you blocked on your first SCOS release or not?
<@jaimelm:fedora.im>
17:04:41
No, we're not blocked. We're doing a bit of a hack though. The nodes are initially FCOS, and then MCO applies SCOS as a layer.
<@dustymabe:matrix.org>
17:05:06
good ol "major downgrade on first boot" :)
<@jaimelm:fedora.im>
17:05:13
ha
<@jaimelm:fedora.im>
17:05:38
Again, these are just initial discussions. No decision necessary now.
<@siosm:matrix.org>
17:05:44
You can also consider migrating to using the FCOS to build SCOS on MOC for now until we find another place to run it
<@siosm:matrix.org>
17:05:56
You can also consider migrating to using the FCOS pipeline to build SCOS on MOC for now until we find another place to run it
<@dustymabe:matrix.org>
17:06:36
TBH, I don't think finding a place to run it will be that hard.. logical answer to me is just another project in the Fedora infra openshift. But of course, that team would have to agree to that
<@dustymabe:matrix.org>
17:07:21
<@dustymabe:matrix.org>
17:07:21
2. in our existing `fedora-coreos-pipeline` project in fedora openshift (not great because separations like credentials aren't clear)
<@dustymabe:matrix.org>
17:07:21
1. centos openshift (is there one?)
<@dustymabe:matrix.org>
17:07:21
3. some other openshift?
<@dustymabe:matrix.org>
17:07:21
other options:
<@jaimelm:fedora.im>
17:08:06
Centos doesn't use OpenShift
<@jbtrystram:matrix.org>
17:08:43
The pipeline can probably run on OKD I guess ?
<@dustymabe:matrix.org>
17:08:47
what's more fuzzy is things like where do we store built arfitacts (s3 bucket access), what accounts do we use to create cloud images, etc..
<@jaimelm:fedora.im>
17:08:59
Right
<@siosm:matrix.org>
17:09:35
which is why I think this should really be owned by the CentOS project
<@siosm:matrix.org>
17:09:57
it should be published under the CentOS projects could accounts, etc.
<@dustymabe:matrix.org>
17:10:02
what does that mean to you travier ?
<@jbtrystram:matrix.org>
17:10:07
Also is SCOS using zincati or a plain OSTree repo like the fedora atomic variants ?
<@siosm:matrix.org>
17:10:09
it should be published under the CentOS project's could accounts, etc.
<@jaimelm:fedora.im>
17:10:15
travier: Can you elaborate?
<@dustymabe:matrix.org>
17:10:27
keep in mind for a long time we built FCOS on an openshift instance managed by the CentOS project :)
<@jaimelm:fedora.im>
17:10:39
and how did that go for you?
<@dustymabe:matrix.org>
17:11:20
was just fine. I'm just pointing out that we mostly just need an openshift instance, doesn't matter who runs it. what matters is how/where we push things (and also things like signing)
<@dustymabe:matrix.org>
17:11:42
<@dustymabe:matrix.org>
17:11:42
> doesn't matter who runs it
<@dustymabe:matrix.org>
17:11:42
well. it does matter. we have to trust them :)
<@jaimelm:fedora.im>
17:12:37
Would an instance of the fedora-coreos-pipeline code be relatively easy to attach different credentials or are there some assumptions hardcoded?
<@jaimelm:fedora.im>
17:12:54
It's Jenkins, right?
<@siosm:matrix.org>
17:13:10
jbtrystram: neither right now. It's only container images as part of an OKD payload
<@jaimelm:fedora.im>
17:13:42
BTW, will FCOS be moving to Konflux?
<@dustymabe:matrix.org>
17:14:17
Though I'm sure we'll find some rough edges
<@dustymabe:matrix.org>
17:14:17
<@dustymabe:matrix.org>
17:14:17
should be. It's kind of built to be reusable.
<@siosm:matrix.org>
17:14:27
It's all 3 parts. We can have the pipeline in Fedora but if we go and push the SCOS images to Fedora repos or cloud accounts then it does not make sense anymore
<@jbtrystram:matrix.org>
17:14:36
travier: i know, but if it's built for standalone use these are to be considered
<@siosm:matrix.org>
17:14:52
It's all 3 parts. We can have the pipeline in Fedora but if we go and push the SCOS container images to Fedora repos or cloud accounts then it does not make sense anymore
<@dustymabe:matrix.org>
17:14:56
travier: indeed. I definitely don't advocate for re-using our Fedora credentials anywhere (like cloud providers)
<@dustymabe:matrix.org>
17:15:18
jbtrystram: good question. I think we kicked the "standalone" question down the road a bit
<@siosm:matrix.org>
17:15:19
jbtrystram: That's the people building SCOS to decide I guess?
<@jaimelm:fedora.im>
17:15:29
Yeah, we wouldn't want to use existing crews.
<@jaimelm:fedora.im>
17:15:42
Yeah, we wouldn't want to use existing creds
<@siosm:matrix.org>
17:15:50
jbtrystram: That's for the people building SCOS to decide I guess?
<@jaimelm:fedora.im>
17:15:55
I think everyone is agreeing on that
<@dustymabe:matrix.org>
17:17:17
Jaime Magiera: running one of these pipelines is a decent amount of work. any chance we'd have someone (or a few) people with some decently high level of engagement to achieve that goal?
<@siosm:matrix.org>
17:17:33
So I'm not opposed to asking Fedora for a namespace to run an SCOS pipeline in, but then we need to find the S3 & cloud credentials to push images to, and those should come from the CentOS project
<@jaimelm:fedora.im>
17:18:26
Yes, that's the plan. We have several people who are interested in digging in deeper. We've just been unable to because of how everything was internal to RH and RH employees.
<@dustymabe:matrix.org>
17:20:18
I think we're mostly done with this topic?
<@jaimelm:fedora.im>
17:20:51
I think so. Initial discussion brought up some good points and things to explore. Thank you.
<@dustymabe:matrix.org>
17:20:54
summary: we're supportive. let's talk more
<@dustymabe:matrix.org>
17:21:48
i see this as mostly us helping guide things along without doing much driving ourselves
<@jaimelm:fedora.im>
17:21:53
Amy is actually in the OKD Working Group. So, I'll ping her on Centos infra stuff.
<@jaimelm:fedora.im>
17:22:26
We'll look at Centos Infra, Fedora Infra, and maybe anything else.
<@jaimelm:fedora.im>
17:22:51
Thanks for everyone's time.
<@jbtrystram:matrix.org>
17:23:15
that's what we hold community meetings for :)
<@jbtrystram:matrix.org>
17:24:13
We don't have other topics scheduled
<@jbtrystram:matrix.org>
17:24:21
!topic open floor
<@jbtrystram:matrix.org>
17:24:36
Does anyone have something else they want to talk about ?
<@jaimelm:fedora.im>
17:24:57
Just to mention that the auto-emails for the meetings still point to the IRC.
<@dustymabe:matrix.org>
17:25:00
!info please try out the `next` stream that has been rebased to Fedora Linux 41 beta content and report issues
<@dustymabe:matrix.org>
17:25:09
<@dustymabe:matrix.org>
17:26:10
I'll fix it
<@dustymabe:matrix.org>
17:27:13
oh fun. the calendar wants me to use `Please use channel@server format!`
<@dustymabe:matrix.org>
17:27:18
i'll have to see what other meetings have done
<@siosm:matrix.org>
17:28:08
You can add a link to the info in the tracker repo instead
<@siosm:matrix.org>
17:29:33
ah, the location
<@siosm:matrix.org>
17:29:37
might not work indeed
<@dustymabe:matrix.org>
17:31:43
let's end this meeting
<@jbtrystram:matrix.org>
17:33:09
yeah we are out of time.
<@jbtrystram:matrix.org>
17:33:12
Thanks everyone !
<@siosm:matrix.org>
17:33:51
Thanks jbtrystram for running, thanks Jaime Magiera for bringing SCOS and thanks all
<@siosm:matrix.org>
17:34:15
!endmeeting
<@jbtrystram:matrix.org>
17:34:17
!endmeeting