Meeting started at 2024-07-24 16:29:50 UTC

The Meeting name is 'fedora_coreos_meeting'

!hi

Timothée Ravier (siosm) - he / him / his

!hi

Hristo Marinov (hricky) - he / him / his

!hi

Adam Piasecki (c4rt0) - he / him / his

!hi

Dusty Mabe (dustymabe) - he / him / his

!hi

Michael Armijo (marmijo)

I'm only here for 30 min so would appreciate if we could start with the Hetzner discussion

!hi gursewak

Gursewak Singh (gursewak)

!hi

Michael Nguyen (mnguyen)

!hi

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

travier: Sure. We can start with the Hetzner discussion!

https://github.com/coreos/fedora-coreos-tracker/issues/1324

Here we have the ticket about it so we can discuss:

hum, we still have to look at actions before

sorry, my request was confusing

Ah, right. I was starting with this so you could participate. But lets see if we have any action items.

Looks there were none from the previous meeting: discussion.fedoraproject.org/t/fedora-coreos-community-meeting-minutes-2024-07-17/126035

Yeah, we don't have any action items.

Let's record this info

now we can go back to the topic at hand :)

!hi aaradhak

Aashish Radhakrishnan (aaradhak)

Nice, thanks!

!topic Platform Request: Hetzner

hum, there is a space before the command in your message so not sure that the bot picked it

OK, now we're good.

The log of last week discussion is in https://meetbot-raw.fedoraproject.org/meeting-1_matrix_fedoraproject-org/2024-07-17/fedora-coreos-meeting.2024-07-17-16.31.log.html#l-1

Docs PR is over there: https://github.com/coreos/fedora-coreos-docs/pull/654

So the questions are:

2. Are we OK with the documentation for this platform (as is or updated to take into account the new images)?

1. Are we OK producing images for Hetzner?

i think if we do `1.` we have to document something.

maybe `2.` is more: "are we OK with referencing and using the `hcloud-image-upload` tool

maybe `2.` is more: "are we OK with referencing and using the `hcloud-image-upload` tool?"

If we produce images for Hetzner, do we have access for basic testing?

we have an account (that I used for testing)

and writing the docs

With a few monthly credits. We could even consider adding release CI for it

(but I'm not advocating for that here)

restating what I said last week.. i'm +1 with creating hetzner images, OR making it convenient (i.e. no copy/paste + run script) to restamp an image via something like `coreos-installer`.

so `1.` 👍

Should we vote for that?

on `2.` - I'm fine with using `hcloud-image-upload` in our docs, but maybe it would be good to reference official hetzner docs (if they exist) for the process, OR link to an open tracking issue for the feature to exist in their official tool

mic check - anyone seeing this?

yes

I'm filling an issue

Yes. Anyone has any other ideas about this topic and want to share?

So let's vote for 1?

proposed: we will start producing Hetzner images for Fedora CoreOS

(it's not great in terms of storage duplication just for a few changed bytes but that would not be the first or last time)

+1

+1

+1

+1

+1

+1

So I think we all agree

!agreed: we will start producing Hetzner images for Fedora CoreOS

!agreed:

It worked now :)

Can we go to the next topic or should we discuss next steps about this one?

I've filled an issue for the tool for the docs

I think we can move on and we'll revisit the docs side of things once we get an answer / I've updated it with the new images?

travier: I would just not block on them giving you an answer on that upstream, but just link it from the docs and ship the docs page to prod :)

YOLO

sgtm :)

I updated the list late last week with the outcome of last week's meeting. There were also a few more changes that came through when I ran the script.

Let's go over the new ones quickly?

marmijo: can Iet you do the next one from the self contained changes? I'll have to go

see you all 👋

thanks

This one looks like it'll probably affect FCOS

I think maybe not.. we don't install libvirt by default so it shouldn't affect stock FCOS

marmijo: and we ship nftables already

Okay, sounds good. Thanks for weighing in!

last one...

given the dependency list I see, I am tempted to say the same thing as the previous one : it should work on FCOS if you install the virt packages

I feel like this one is something that depends

all existing uses of FCOS today should continue fine so there is nothing for us to do to keep the features we have today

but, if we want to support this new type of confidential compute (i.e. FCOS guests), we'd probably need to do some things

maybe we just open a feature request in our tracker for the new feature?

i.e. point to the change request and say get this working with FCOS as a guest

I wouldn't be surprised if this isn't wanted for RHCOS too

I would be surprised if this isn't wanted for RHCOS too

Okay, I'll open that issue.

dustymabe: I am not sure what makes you say that this needs change on the guest ?

In that proposal i only see changes for the host

dustymabe: I am not sure what makes you say that this needs changes on the guest side of things ?

jbtrystram: I assue the "guest attestation" will require tools inside the guest to perform that attestation

jbtrystram: I assume the "guest attestation" will require tools inside the guest to perform that attestation

but I could be wrong!

either way - if it does just work out of the box maybe we should add a test for it?

so either way I see there being work for us to do

that guest attestation thing isn't really precise right now, but I guess we can monitor that yes

Those are all of the new changes 👍️

So, we are almost done with time. Can we finish this topic?

So we can go to open floor :)

yup, you're good to move on!

!topic Open Floor

If someone has anything to share in those last minutes be free to do it

Well, lets ending the meeting!

Thank you all!