Meeting started at 2025-09-10 17:00:08 UTC

The Meeting name is 'fedora-server'

I'll post the agenda in 2-3 minutes.

As usual, let's wait a moment for everybody to show up.

Hello, Peter Boy (ServerWG, Docs) !

!hello

Emmanuel Seyman (eseyman) - he / him / his

Welcome Emmanuel!

Thanks for writing to the tomcat folks!

I just got back from work and have read their reply.

yeah I read that, too. Well, a case about "scope" and "assignment"

And usually, a reverse proxy uses localhost, not really a "network" access.

And the default configuration is using localhost, too

in this case, localhost is network access

Yes, but no public access and no security thread

Yes, but no public access and no security threat

Given that you're running a reverse proxy, "no public access" seems a bit of a stretch

OK, yeah, but there is no direct public access to localhost.

And is someone compromises the system, you may have more urgent issues as localhost. :-)

anyway, will reply later tonight

OK, let's see here. We are just the two of us and nirik lurking as usual. I think, we short have a short look at the topcis, nevertheless.

Let's do that

Regarding the action, nothing new. See

And no announcements from me.

Anything else for the record?

No

Tracking issue: !link https://pagure.io/fedora-server/issue/169

At the moment there is nothing to do about this. But probably any questions?

None, I see. So we are waiting for abbra 's proposal.

Nope

Yes, I trust abbra 's judgement here

I need to make sure someone could do this work first. As Adam explained, there are some unsolved issues with the infra part. Once we are able to add the automation, we can expand the release criteria

OK. How long cat we wait, or with other words, how essential is that?

But we can deal with that when nothing is in sight for F44/45 yet.

Traxking issue: !link https://pagure.io/fedora-server/issue/166

Emmanuel Seyman: what's your first idea about that? You proposed some time ago to remove the SBC part from our main installation section.

So it was originally your idea :-)

Did I?

Yes!!!

But you may have changed your mind in the meantime.

And I suspect the tracking issue is not that one

oh

It is https://pagure.io/fedora-server/issue/168

Much better

yeah, copy&paste

My issue is that "SBC board" and "underpowered" are pretty much joined at the hip, in my mind

If CPU power is a concern, I suspect you're going to quickly disregard ARM SBC boards

Yeah, but have a look at the new generation of Rockchip 3586 chips. They are not that much slower as an Intel n100 box. It's a big step forward.

And you can use a "real" bios and Anconda install.

But yes, there is a gap. So i think it is important, to draw a clear separation line.

but separating the "write to micro-SD card" and "do the actual install" use-cases seems a good idea

We should/must emphasize that a suitable storage medium must be on board. The current Raspberry Pi 4 is a no-go.

Well, what would be the alternative? Should we drop ARM SBC at all?

We'll have a similiar question with the Risc V boards. I testet 2 of them (which currently the development boards of the SIG).

They are generally better suited for server. But they are SBCs

I would hate to lose users simply because we don't want to maintain it

So I take it as agreeing to follow the proposed path, basically. details to be discussed.?

Yes

Regarding to lose users: Currently we can't F43 install at all. :-)

So I'll finetune the text for a more intensive discussion later.

So I'll finetune the text a bit for a more intensive discussion later.

Tracking issue: !link https://pagure.io/fedora-server/issue/164

The beta candidate is out today. On Thursday there is the go/no-go meeting.

I started testing today with the SBC installation, which didn't work with the branched and rawhide version.

Still haven't found the time to test installation. Spent the weekend releasing security updates

I still doesn't work: ee: https://bugzilla.redhat.com/show_bug.cgi?id=2391231

Regarding testing: I'll be able to test our hardware installation part. Would be nice, if someone else could test the VM installations.

In my previous test I didn't find a serious issue. That's why I'm optimistic.

That bug does look relatively easy to fix, if Peter R's comments are true

Peter's provided an update to Arm tools, but it didn't fix the arm-installer-issue.

And the ARM SIG is currently not in the best state.

I will try tomorrowx

I will try tomorrow

👍️

Maybe you can compare the F43 arm-image-installer with the F40 / F41 version. Those worked as far as I remember

I don't know about F42. I think, we never tested that one.

Might as well try F41, F42 and F43 inthat order

Might as well try F41, F42 and F43 in that order

Ok, yes. indeed the best way.

Another issue is, the Beta does not include the minimal version. We need that for installation of devices with a soldered SPI module.

It's really bad with the ARM devices this time.

OK, let's proceed.

I think we should skip again and wait for Paul Maconi (Aggraxis)

And probably discuss in our Matrix room.

Let's

OK.

I'm still hoping the new forge will help here

topic 6. Open Floor

Anything here?

My question, we soon will have the switch to winter time.

any news on the PR against the httpd package?

Can we follow the US timinn again? So we 2 have to change our daily routine for 2 meetings.

This would make the meeting earlier, then ?

Regarding httpd: I have prepared the PR. I'm currentl looking into the build file for an opportunity to create the additional directory.

Yeah, i suppose earlier, 1 hour.

Back to httpd: When I made the PR, I'll sent a mail with the link.

I've had a few ideas since that I will email you about today or tomorrow

re US time : I can make it work

OK, so I will adjust the Fedora calendar accordingly.

And wait for the mail.

Anything else?

No, I'm good

Well, obviously not. Someone else lurking around with question or comment?

I think we're good (and I want to have dinner)

OK