#meeting-1:fedoraproject.org log

<@siosm:matrix.org>

16:30:47

!startmeeting fedora_coreos_meeting

<@meetbot:fedora.im>

16:30:52

Meeting started at 2024-08-21 16:30:47 UTC

<@meetbot:fedora.im>

16:30:52

The Meeting name is 'fedora_coreos_meeting'

<@siosm:matrix.org>

16:30:55

!topic roll call

<@dustymabe:matrix.org>

16:31:02

!hi

<@zodbot:fedora.im>

16:31:18

Dusty Mabe (dustymabe) - he / him / his

<@siosm:matrix.org>

16:31:24

!hi

<@zodbot:fedora.im>

16:31:29

Timothée Ravier (siosm) - he / him / his

<@aaradhak:matrix.org>

16:31:30

!hi aaradhak

<@zodbot:fedora.im>

16:31:31

Aashish Radhakrishnan (aaradhak)

<@ydesouza:fedora.im>

16:31:35

!hi

<@zodbot:fedora.im>

16:31:36

Yasmin Valim de Souza (ydesouza)

<@hricky:fedora.im>

16:31:43

!hi

<@zodbot:fedora.im>

16:31:44

Hristo Marinov (hricky) - he / him / his

<@marmijo:fedora.im>

16:32:07

!hi

<@zodbot:fedora.im>

16:32:08

Michael Armijo (marmijo)

<@jlebon:fedora.im>

16:32:47

!hi

<@zodbot:fedora.im>

16:32:48

None (jlebon)

<@ravanelli:matrix.org>

16:34:06

!hi ravanelli

<@zodbot:fedora.im>

16:34:08

Renata Ravanelli (ravanelli)

<@siosm:matrix.org>

16:34:12

Alright, let's get started!

<@siosm:matrix.org>

16:34:18

!topic Action items from last meeting

<@jbtrystram:matrix.org>

16:34:19

!hi

<@zodbot:fedora.im>

16:34:22

Jean-Baptiste Trystram (jbtrystram) - he / him / his

<@siosm:matrix.org>

16:35:04

!link https://discussion.fedoraproject.org/t/fedora-coreos-community-meeting-minutes-2024-08-14/128688

<@siosm:matrix.org>

16:35:11

I see no action items from last time

<@siosm:matrix.org>

16:35:43

Let's go to today's topics

<@gurssing:matrix.org>

16:35:44

!hi gursewak

<@zodbot:fedora.im>

16:35:46

Gursewak Singh (gursewak)

<@siosm:matrix.org>

16:35:50

!topic tracker: Fedora 41 changes considerations

<@siosm:matrix.org>

16:35:54

!link https://github.com/coreos/fedora-coreos-tracker/issues/1714

<@siosm:matrix.org>

16:36:30

Now that Fedora 41 has been branched, there should not be that much change to this list, normally only removals / deferred changes

<@siosm:matrix.org>

16:36:52

We have 2 to review

<@siosm:matrix.org>

16:37:14

Thanks to marmijo for keeping the changelist updated!

<@marmijo:fedora.im>

16:37:24

you're welcome!

<@siosm:matrix.org>

16:38:09

230. ROCm 6.2

<@siosm:matrix.org>

16:38:20

- 230. ROCm 6.2

<@siosm:matrix.org>

16:38:26

!link https://fedoraproject.org/wiki/Changes/ROCm6.2

<@siosm:matrix.org>

16:38:47

I don't think we ship this package so that should not impact us?

<@jlebon:fedora.im>

16:39:43

👍️

<@siosm:matrix.org>

16:40:51

- 231 Enabling composefs by default for Atomic Desktops, CoreOS and IoT

<@siosm:matrix.org>

16:40:57

!link https://fedoraproject.org/wiki/Changes/ComposefsAtomicCoreOSIoT

<@siosm:matrix.org>

16:41:19

!link https://github.com/coreos/fedora-coreos-tracker/issues/1718

<@siosm:matrix.org>

16:41:49

We are driving this one so we're aware of it.

<@siosm:matrix.org>

16:42:04

So I think that's all for this list?

<@siosm:matrix.org>

16:42:35

We can talk of the status of composefs in the next topic (status of branched fcos)

<@siosm:matrix.org>

16:42:39

We can talk of the status of composefs in the next topic (status of branched FCOS)

<@marmijo:fedora.im>

16:43:47

Sounds good. Fedora 41 just branched, so it might be worth checking in on the status of the changes that will affect FCOS.

<@siosm:matrix.org>

16:44:25

!topic tracker: Rebase onto Fedora 41

<@siosm:matrix.org>

16:44:29

!link https://github.com/coreos/fedora-coreos-tracker/issues/1695

<@siosm:matrix.org>

16:44:36

!undo

<@siosm:matrix.org>

16:44:44

!topic tracker: Rebase onto Fedora 41

<@siosm:matrix.org>

16:44:51

!link https://github.com/coreos/fedora-coreos-tracker/issues/1695

<@siosm:matrix.org>

16:45:38

Let's talk about the state of the branched stream for FCOS, what are the priority and the state of our changes.

<@jlebon:fedora.im>

16:46:17

i have one item to start the conversation with

<@jlebon:fedora.im>

16:47:33

currently, we're pinning to systemd v255. f41 is shipping with v256.

<@jlebon:fedora.im>

16:47:33

https://github.com/coreos/fedora-coreos-tracker/issues/1775

<@jlebon:fedora.im>

16:47:33

the issue we're hitting is that v256 somehow is breaking SSH. i started looking more deeply into this to fix ASAP.

<@jlebon:fedora.im>

16:48:15

there's a chance that once we catch up to v256, that'll unmask new issues since that'll be our first exposure to that version

<@jbtrystram:matrix.org>

16:48:54

systemd255 have been pinned for long now

<@jlebon:fedora.im>

16:49:02

this really should've been looked at much earlier, but for a while it was entangled with SELinux issues and it wasn't clear that it was an independent issue

<@marmijo:fedora.im>

16:49:12

definitely. It'll be important for us to try to resolve those new issues quickly

<@siosm:matrix.org>

16:51:39

Agree that systemd 256 should be the focus.

<@siosm:matrix.org>

16:52:12

We had planned to do composefs (https://github.com/coreos/fedora-coreos-tracker/issues/1718) but it's blocked on an issue with kdump.

<@dustymabe:matrix.org>

16:52:30

is silverblue shipping the new systemd? can those systems be SSH into?

<@dustymabe:matrix.org>

16:52:35

i.e. how is this only affecting FCOS?

<@siosm:matrix.org>

16:52:45

For the Atomic Desktops, it's likely we'll drop composefs from F41 as we can not complete the work in time for F41. Might as well drop it from FCOS for F41

<@jbtrystram:matrix.org>

16:53:12

travier: isn't it fixed by now ? cgwalters submitted a patch to kdump. let me check

<@siosm:matrix.org>

16:53:29

dustymabe: That's a good question. I haven't had the time to test with F41 Atomic Desktops yet

<@jlebon:fedora.im>

16:53:31

dustymabe: AFAIK only FCOS

<@jlebon:fedora.im>

16:55:13

travier: is the kdump issue the only blocker?

<@siosm:matrix.org>

16:55:35

For FCOS yes AFAIK. For the Atomic Deskops we have other issues

<@jlebon:fedora.im>

16:56:32

ok. we should really try to land in FCOS. could we document that if you want to use kdump, for now you have to disable composefs?

<@jbtrystram:matrix.org>

16:56:40

The kdump issue is fixed, but the PR is still opened. It will probably be merged soon as people get back from holiday ? I'm hopeful :)

<@jlebon:fedora.im>

16:57:19

ok. we should really try to land in FCOS. could we document that if you want to use kdump, for now you have to disable composefs?

<@jlebon:fedora.im>

16:57:19

(which I think is also something that the podman team will want to do that we need to give a heads up on)

<@siosm:matrix.org>

16:58:07

The PR does not look ready

<@jbtrystram:matrix.org>

16:58:26

One workaround is to set kdump to a remote target , as the composeFS issue arise only when the dump target is a local filesystem

<@siosm:matrix.org>

16:58:39

I don't understand how the kdump issue can be "fixed" in RHEL without a PR merged in CentOS Stream/Fedora/upstream

<@dustymabe:matrix.org>

16:59:53

we can maybe revisit the composefs issue in a week - maybe the PR will have merged by then

<@dustymabe:matrix.org>

17:00:14

if it doesn't make it in by beta freeze then it will be harder to get in and we can probably punt to F41

<@dustymabe:matrix.org>

17:00:32

assuming the kdump issue is the only remaining issue?

<@jbtrystram:matrix.org>

17:00:55

travier: it's probably patched in git-diff, not in kdump upstream yet (i'm looking)

<@jbtrystram:matrix.org>

17:01:05

*git-dist

<@jbtrystram:matrix.org>

17:02:00

yeah that's it : https://gitlab.com/redhat/centos-stream/rpms/kexec-tools/-/commit/099aead590608cd9edb49acf73873b03b88576b6 and https://gitlab.com/redhat/centos-stream/rpms/kexec-tools/-/commit/be56205d06d05ce7d1d6313cf5141a34bec597e2

<@jlebon:fedora.im>

17:02:02

we could also let it get to next, but not promote it to testing at GA if we don't feel it's ready yet

<@jlebon:fedora.im>

17:03:23

at least for FCOS, kdump doesn't feel like it's worth hard blocking on if it's really the only issue IMO

<@dustymabe:matrix.org>

17:04:09

I agree, but it would be nice to make sure it does pass once the kdump fix is in (i.e. some followup we do there to ensure it lands and works)

<@jlebon:fedora.im>

17:04:13

but it sounds like this should be resolved soon, so maybe it's moot

<@siosm:matrix.org>

17:04:53

I did a quick check for the added/removed packages in https://github.com/coreos/fedora-coreos-tracker/issues/1695#issuecomment-2298815019

<@siosm:matrix.org>

17:05:35

we should investigate at least the pigz & polkit-pkla-compat removals

<@marmijo:fedora.im>

17:06:16

Should I open a tracker issue for package additions/removals? We dont have one yet for F41.

<@dustymabe:matrix.org>

17:06:28

that would be great!

<@jlebon:fedora.im>

17:06:59

looks like we lost a bunch more wifi firmwares

<@dustymabe:matrix.org>

17:07:46

Jonathan Lebon: on purpose, right?

<@jlebon:fedora.im>

17:08:50

hmm, the addition of tini-static looks suspect. or at least i would've expected also a removal of catatonit assuming it comes from podman switching over

<@jlebon:fedora.im>

17:09:38

dustymabe: yup, definitely. matches https://github.com/coreos/fedora-coreos-tracker/issues/1575

<@marmijo:fedora.im>

17:10:08

!link https://github.com/coreos/fedora-coreos-tracker/issues/1785

<@siosm:matrix.org>

17:12:33

DNF5 is https://github.com/coreos/fedora-coreos-tracker/issues/1687

<@siosm:matrix.org>

17:12:37

!link https://github.com/coreos/fedora-coreos-tracker/issues/1687

<@jbtrystram:matrix.org>

17:13:17

qed-firmware was added after it was split out from linux-firmware : https://github.com/coreos/fedora-coreos-tracker/issues/1746

<@jlebon:fedora.im>

17:14:23

ahh right, i think we said we would re-discuss the dnf5 item

<@siosm:matrix.org>

17:14:36

How do we feel about dnf5 in FCOS? It's mostly got some fixes but is it going to be useful?

<@siosm:matrix.org>

17:15:10

hum, I don't even know if those "warning message" fixes landed in dnf5 at all

<@jlebon:fedora.im>

17:15:34

i don't think it did yet

<@siosm:matrix.org>

17:15:36

dnf4 still need https://github.com/rpm-software-management/dnf/pull/2121 but that's not for us

<@siosm:matrix.org>

17:16:16

I see no PR mentioning bootc in dnf5: https://github.com/rpm-software-management/dnf5/pulls?q=sort%3Aupdated-desc+is%3Apr+bootc

<@jlebon:fedora.im>

17:17:04

there's still a lot to iron out re. the layering & updates story. until then, i think it's useful for anyone choosing their own FCOS layering adventure, or locally in usroverlay mode. and of course for testing things

<@siosm:matrix.org>

17:17:29

Side note: For the Atomic Desktops, we'll only include bootc & dnf5 in the container version, not in the classic ostree one as they won't really be useable there.

<@jlebon:fedora.im>

17:18:25

really had hoped the warnings at least would've landed by now

<@jlebon:fedora.im>

17:19:29

travier: makes sense re. atomic desktops. i guess we're a bit in between. we're still on ostree but also publish container images

<@dustymabe:matrix.org>

17:20:17

i guess dnf5 at least doesn't pull in python?

<@siosm:matrix.org>

17:20:25

(We're close to time to unless we got something else urgent, I'll move to open floor)

<@jlebon:fedora.im>

17:20:35

(that's another one that could make sense to leave in next until we feel it's ready to put in testing)

<@jlebon:fedora.im>

17:20:41

dustymabe: indeed

<@dustymabe:matrix.org>

17:21:10

Jonathan Lebon: though note, that the closer `next` is to what we will promote to `testing` for the 40->41 switch, the better

<@dustymabe:matrix.org>

17:21:21

Jonathan Lebon: though note that the closer `next` is to what we will promote to `testing` for the 40->41 switch, the better

<@jlebon:fedora.im>

17:21:32

dustymabe: agree

<@jlebon:fedora.im>

17:22:10

but next was also built for experimental things.

<@jlebon:fedora.im>

17:22:10

if we go that way, we could wait until after GA to land it in next

<@siosm:matrix.org>

17:23:26

!topic Open Floor

<@dustymabe:matrix.org>

17:26:03

👋 - i'll be a little more around now

<@siosm:matrix.org>

17:26:16

Thanks to marmijo for setting up the branched stream for F41!

<@marmijo:fedora.im>

17:26:31

Of course!

<@siosm:matrix.org>

17:27:16

and thanks to Yasmin Valim de Souza for the releases last week

<@jlebon:fedora.im>

17:27:22

marmijo++

<@jlebon:fedora.im>

17:27:22

yes, thank you so much marmijo for working on this!

<@dustymabe:matrix.org>

17:27:46

🎉

<@siosm:matrix.org>

17:30:44

!endmeeting