Meeting started at 2024-06-05 16:38:07 UTC

The Meeting name is 'fedora_coreos_meeting'

!hi

Yasmin Valim de Souza (ydesouza)

!hi

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

!hi

Nishant Mishra (rico0098)

!hi

Hristo Marinov (hricky) - he / him / his

!hi aaradhak

I'll start running the meeting until gursewak show up :)

Aashish Radhakrishnan (aaradhak)

.hi gursewak

!hi

Adam Piasecki (c4rt0) - he / him / his

!hi

Nishant Mishra (rico0098)

!hi

!hi

Joseph Marrero (jmarrero)

Dusty Mabe (dustymabe) - he / him / his

.hi gursewak

!hi

BTW is the !topic roll call really needed anymore ?

Michael Armijo (marmijo)

zodbot seems to pick people just fine

ah gursewak :) I'll let you take over :)

!hi gursewak

Gursewak Singh (gursewak)

I guess it gives us some time to copy commands

Ouch, I hit enter before completing the sentence... let's leave it at that :D

I see no action items from last meeting.

jbtrystram: do you want to introduce this?

yeah :) The NFS module is explicitly excluded from the intiramfs, as we don't support root over NFS. However, mounting NFS in the initramfs can be used for other reason, such as writing kernel memory dumps to remote targets, which is done in kdump own initramfs

when kdump builds it's initramfs, it copies the existing one. Which cause NFS remotes to fail because of the missing module

I was thinking it's worth including the module for that use case. it would requires some kargs handling to error out on an NFS-root however

!hi jasonbrooks

Jason Brooks (jasonbrooks) - he / him / his

shouldn't kdump know to ensure the NFS bit is in its initramfs if it sees that an NFS target is configured? and similarly for other targets

There is probably some edge case that i won't think about, so I thought I'd bring that up

it seems awkward to add it to the official initramfs just for kdump

Jonathan Lebon: there is the `extra_modules` directive in kdump.conf which should work

I haven't tested though

jbtrystram: might be worth discussing this with the kdump maintainers

I think what Jonathan Lebon is trying to say is why doesn't kdump detect and do the right thing here?

Maybe just a doc update will b enough :)

i.e. the fix here isn't in our stack, but upstream

I'll ask upstream then :)

we can move on i think :)

alright.

i think for this one, it might be easier to do a separate sync where we go over the list, and then only bring up what needs discussions to the next community meeting

That works for me. Since dusty is on leave for now, Ash and I will be managing the list located in the attached tracker issue. We'll update the list once a week so we can discuss the items throughout the next few months in this meeting (like we normally do).

For now it looks like most items will either not affect Fedora CoreOS, or there are items that were deferred from F40 that we may have already been discussing.

marmijo: can you find a good timeslot and announce it in the ticket?

and make sure at least travier and I join :)

Absolutely. I'll schedule something with the usual players, but if anyone wants to join, just let me know

thanks for taking care of keeping the list up to date!

Of course, I'll update it again this week so we can discuss the newest items

marmijo: ++

dustymabe gave a cookie to marmijo. They now have 2 cookies, 1 of which were obtained in the Fedora 40 release cycle

Sounds good. That's pretty much what we have for the issues. Moving on to open floor unless there's more to add.

to clarify, what I mean is: let's put the meeting details in the tracker issue itself to make sure that anyone who wants to join can

Sounds good. I'll be sure to do that

i have a few items :)

is this the coreos meeting?

I just realized

everything went over my head my first meeting since F37

first: dustymabe since you're here, two items for you:

- can you give jbtrystram an account for the Fedora AWS community account?

- VexxHost is hitting quota issues; I think you've had to do manual stuff in the past. is that documented somewhere/does someone else know what needs to happen there?

Nishant Mishra: yes, it is :)

Jonathan Lebon: for vexxhost I think someone working on https://github.com/coreos/fedora-coreos-pipeline/issues/900 would go a long way maybe?

dustymabe: have you been manually deleting stuff whenever we hit our quota?

oof, I forgot that was assigned to me. dustymabe I'll try to work on that soon.

Jonathan Lebon: yeah, occasionally - it's not often, but it does happen: https://github.com/coreos/fedora-coreos-pipeline/issues/889

ok fun

I'll try to clean things up right now

if we are hitting keypair limits then https://github.com/coreos/coreos-assembler/issues/3550 might be important

right now, it's space issues

marmijo: ++

jlebon gave a cookie to marmijo. They now have 3 cookies, 2 of which were obtained in the Fedora 40 release cycle

Jonathan Lebon: then running `ore openstack gc` with the appropriate creds in the pipeline should work then

just replay the job and get it to sleep and pop into the pod to get in the right env and run it ?

vexxhost infra sometimes can flake, so the request to delete the resources might not 100% work, if it doesn't clear up I can look at the web interface too

the other item I had was: is anyone interested in bumping our various container images to f40? it's the last section in https://github.com/coreos/fedora-coreos-tracker/issues/1674

we can also split it up. i think apiaseck did the bulk of them last cycle

I can look into that.

I was typing just that. Let me know if you need a hand ash

apiaseck ++

ash ++

awesome!

sure, will check with you on this.

anyone else have anything they want to bring up?

Will give couple of minutes before I end the meeting:)

:) - I guess I should open a package inclusion request for firewalld

need to run - thanks for running the meeting gursewak

dustymabe: I can try to do that if you don't mind.

Hristo Marinov: Dusty's on vacation so I am pretty sure you can do that.