#meeting:fedoraproject.org log

<@salimma:fedora.im>

15:01:38

!startmeeting CentOS Hyperscale SIG

<@meetbot:fedora.im>

15:01:40

Meeting started at 2025-03-26 15:01:38 UTC

<@meetbot:fedora.im>

15:01:40

The Meeting name is 'CentOS Hyperscale SIG'

<@salimma:fedora.im>

15:01:45

!topic Roll Call

<@amoloney:fedora.im>

15:02:38

Aw no, we really need to fix fedocal

<@amoloney:fedora.im>

15:02:51

Council meeting can happen in the council chat room

<@amoloney:fedora.im>

15:03:36

For those wanting to join the council meeting, please head to #meeting-1:fedoraproject.org

<@amoloney:fedora.im>

15:03:44

sorry for the disruption Michel Lind UTC-6 !

<@salimma:fedora.im>

15:04:12

np Aoife Moloney

<@salimma:fedora.im>

15:07:58

let's give it 3 more minutes then if nobody shows up I'm going to cancel the meeting

<@ngompa:fedora.im>

15:09:27

!hi

<@zodbot:fedora.im>

15:09:29

Neal Gompa (ngompa) - he / him / his

<@ngompa:fedora.im>

15:09:31

ffs netsplit

<@salimma:fedora.im>

15:09:37

about time, I was going to cancel

<@ngompa:fedora.im>

15:09:42

none of my messages on my main account were going through

<@salimma:fedora.im>

15:09:47

Matrix broke again? jeesus

<@ngompa:fedora.im>

15:09:48

I noticed because zodbot wasn't responding

<@ngompa:fedora.im>

15:10:13

both Daan De Meyer and rcolebaugh are here too

<@salimma:fedora.im>

15:10:17

and your other matrix account ... is matrix.org right? yikes if they can't keep federation between that and EMS up

<@ngompa:fedora.im>

15:10:22

yes

<@ngompa:fedora.im>

15:10:53

something is wrong with the fedoraproject EMS instance

<@ngompa:fedora.im>

15:11:13

because I have not encountered this problem elsewhere

<@salimma:fedora.im>

15:12:02

sigh that's even worse

<@ngompa:fedora.im>

15:12:10

so the status sync is working, just not my messages

<@salimma:fedora.im>

15:12:42

let me test with my other EMS account

<@salimma:fedora.im>

15:14:37

of course I quit most rooms on the other account because otherwise things are so slow on mobile

<@michel:one.ems.host>

15:15:45

Test

<@salimma:fedora.im>

15:16:04

ok, so Element One -> Fedora's EMS instance works at least

<@ngompa:fedora.im>

15:16:22

I'm reloading Element after upgrading the flatpak

<@ngompa:fedora.im>

15:16:25

maybe that fixes it

<@salimma:fedora.im>

15:18:03

let's wait a few minutes then in case more people will show up

<@salimma:fedora.im>

15:18:08

rcolebaugh: is Vishal joining us today?

<@ngompa:fedora.im>

15:18:38

he was trying

<@ngompa:fedora.im>

15:18:42

I saw him typing

<@ngompa:fedora.im>

15:20:22

he's here as `@vishalmishra434:matrix.org`

<@ngompa:fedora.im>

15:20:34

@vishalmishra434:matrix.org has been typing and just not making it through

<@salimma:fedora.im>

15:21:37

oh god is it the matrix outage again

<@salimma:fedora.im>

15:21:53

so yeah matrix.org <-> EMS is split brained then. that's... really not cool

<@conan_kudo:matrix.org>

15:21:57

the problem is that Michel Lind UTC-6 is on fedora.im and can't see it

<@daandemeyer/:matrix.org>

15:22:13

Time to move to whatsapp

<@vishalmishra434:matrix.org>

15:22:17

Any suggestions on what I should do?

<@rcolebaugh:matrix.org>

15:22:24

Aww that is so annoying

<@salimma:fedora.im>

15:22:26

let's ask Vishal to get a Fedora account and get on fedora.im instead, that's the only one that is guaranteed to work

<@salimma:fedora.im>

15:22:50

same with Raymond who's also using a matrix.org account

<@salimma:fedora.im>

15:22:57

meanwhile let me kvetch at the Element people

<@ngompa:fedora.im>

15:23:01

he can't

<@ngompa:fedora.im>

15:23:13

there are no more slots left on the fedora.im server for new users

<@ngompa:fedora.im>

15:23:22

it was maxed out earlier this month

<@ngompa:fedora.im>

15:24:52

See https://mastodon.social/@andyprice/114202658571224733

<@nirik:matrix.scrye.com>

15:26:37

That was fixed several days ago

<@ngompa:fedora.im>

15:26:59

oh hey I see Vishal Mishra's message

<@ngompa:fedora.im>

15:27:01

and rcolebaugh

<@ngompa:fedora.im>

15:27:19

it seems like it's slowly syncing

<@salimma:fedora.im>

15:27:27

FFS

<@salimma:fedora.im>

15:27:34

OK I can see this now

<@vishalmishra434:fedora.im>

15:28:08

!hi

<@zodbot:fedora.im>

15:28:09

Vishal Mishra (vishalmishra434)

<@salimma:fedora.im>

15:28:14

the problem is the *bot* is in this room so if I can't see it it can't either

<@salimma:fedora.im>

15:28:43

ok, so everyone is here now I guess

<@salimma:fedora.im>

15:28:56

let's continue because we're almost halfway in

<@ngompa:fedora.im>

15:28:59

it only took... *checks notes* 28 minutes

<@ngompa:fedora.im>

15:29:14

we'll just speed run this meeting

<@salimma:fedora.im>

15:29:17

yeah. urgh. I did just complain to the Element one and EMS room so I wonder if that helped. but too soon

<@salimma:fedora.im>

15:29:21

!topic Followups

<@salimma:fedora.im>

15:29:34

anyone? if not I have one

<@ngompa:fedora.im>

15:29:46

I am beating my head into the ground trying to get hyperscale 10 composes work

<@ngompa:fedora.im>

15:29:48

I am beating my head into the ground trying to get hyperscale 10 composes working

<@nhanlon:beeper.com>

15:30:11

That's no fun

<@nhanlon:beeper.com>

15:30:15

(also hi)

<@salimma:fedora.im>

15:30:21

hi Neil

<@nhanlon:beeper.com>

15:30:34

just lost as usual, so poking in 🙂

<@ngompa:fedora.im>

15:30:57

I'll probably do a kernel rebase and talk to Daan De Meyer about an selinux-policy rebase and see if that helps

<@ngompa:fedora.im>

15:31:10

because I'm at my wits end for why this is broken

<@salimma:fedora.im>

15:31:25

what is broken exactly? kernel related I guess

<@ngompa:fedora.im>

15:31:33

the kernel doesn't boot at all

<@ngompa:fedora.im>

15:31:36

which I have no idea why

<@salimma:fedora.im>

15:31:39

!info ngompa is still working on getting Hyperscale 10 composes but is stuck

<@salimma:fedora.im>

15:31:54

can we... triage by cheating and using the Fedora kernel for now and see if that works?

<@ngompa:fedora.im>

15:31:59

and the selinux-policy seems to be somehow broken in such a way that image builds fail with undefined label errors

<@salimma:fedora.im>

15:32:17

oof... uh, that's not good

<@ngompa:fedora.im>

15:32:39

I've never seen these failures before

<@nhanlon:beeper.com>

15:32:44

**sends positive vibes to Neal**

<@ngompa:fedora.im>

15:32:47

not on Fedora, and not on CS9 either

<@salimma:fedora.im>

15:34:54

can we turn off selinux in the compose for now? not ideal, but at least separate the kernel issue from the selinux issue (and who knows if we're really (un)lucky they're one and the same

<@ngompa:fedora.im>

15:35:19

if a rebase of the kernel fixes its bootable-ness, sure

<@salimma:fedora.im>

15:35:53

yeah, just unblock ourselves as much as possible - as long as we document what's not working that should be fine

<@salimma:fedora.im>

15:36:00

plus you can always turn it back on later and relabel

<@salimma:fedora.im>

15:36:08

let's move on

<@ngompa:fedora.im>

15:36:16

what's stupid is that it works fine if you go from vanilla to hs10

<@salimma:fedora.im>

15:36:38

oh, so converting to hs10 works - you just don't get to keep a btrfs filesystem

<@salimma:fedora.im>

15:36:49

wait, even the hs kernel work? how

<@salimma:fedora.im>

15:36:57

anyway... given time is short today we probably should mvoe on

<@ngompa:fedora.im>

15:37:12

yes, let's

<@salimma:fedora.im>

15:37:24

!info Proposed Updates SIG is up, any update we're trying to get into CentOS Stream proper should be built there and not in Hyperscale anymore

<@ngompa:fedora.im>

15:37:25

we can discuss my pain in #centos-hyperscale:fedoraproject.org later

<@ngompa:fedora.im>

15:37:54

I'm happy to see that we have proposed updates sig going now

<@ngompa:fedora.im>

15:38:14

though I'm not sure if this means I'll look less active in hyperscale now? that is a lot of the stuff I do 😓

<@ngompa:fedora.im>

15:38:17

though I'm not sure if this means I'll look less active in hyperscale now? that is a lot of the stuff I do 😅

<@salimma:fedora.im>

15:38:19

we don't have c10s yet - I'm going to put up Neal's mutter MR there to start it - and we still need docs and then we can properly announce it. It's also a talk in the CentOS Showcase on Apr 7 and we will be at the community pavillion at RH summit

<@salimma:fedora.im>

15:38:37

eh, our total activity will be up. Hyperscale activity going down is actually a plus I guess

<@salimma:fedora.im>

15:38:59

officially we only have Jonathan, me and Davide now in the CPU SIG, but once docs are up we should start onboarding members

<@salimma:fedora.im>

15:39:30

and... unless anyone has questions let's move on

<@ngompa:fedora.im>

15:39:41

none from me, let's keep moving

<@salimma:fedora.im>

15:39:46

!topic Announcements

<@salimma:fedora.im>

15:40:04

!info CentOS Showcase on April 7. CFP is really, really closed this time I guess but if you want to register (free online event) - is there an outage between matrix.org and EMS? messages from matrix.org can't be seen from EMS, I only see "... is typing"

<@salimma:fedora.im>

15:40:12

oops

<@salimma:fedora.im>

15:40:23

!link https://www.centos.org/events/showcase-2025-04

<@ngompa:fedora.im>

15:40:46

lol

<@ngompa:fedora.im>

15:41:05

schedule for the showcase isn't up yet?

<@salimma:fedora.im>

15:41:08

if you really want a talk in we can talk to Shaun and there might still be spots

<@salimma:fedora.im>

15:41:36

I just got my talk in on Monday, and Shaun said if there's not enough submissions it might get canceled, so ... I'm guessing either will happen soon

<@salimma:fedora.im>

15:41:59

(the CFP closed unexpectedly early last Friday, which I think did not help getting talks in :P)

<@salimma:fedora.im>

15:42:08

any other announcement?

<@ngompa:fedora.im>

15:42:23

the Summit stuff?

<@salimma:fedora.im>

15:42:25

we can speedrun tickets so we can get to membership and not keep Vishal Mishra waiting

<@ngompa:fedora.im>

15:42:27

the schedule is now out

<@salimma:fedora.im>

15:42:38

yeah if you have the link go for it and announce it here

<@ngompa:fedora.im>

15:43:51

!info Davide and Neal will be giving a talk about CentOS Hyperscale Asahi Remix at Red Hat Summit on May 19 in Boston

<@ngompa:fedora.im>

15:43:53

!link https://events.experiences.redhat.com/widget/redhat/sum25/SessionCatalog2025/session/1731519631980001Xort

<@salimma:fedora.im>

15:44:25

nice

<@salimma:fedora.im>

15:44:41

(my talk is not Hyperscale related so I won't plug it today)

<@salimma:fedora.im>

15:44:56

well ... it might be since we're using Hyperscale to incubate GNOME

<@ngompa:fedora.im>

15:45:02

it definitely is :)

<@salimma:fedora.im>

15:45:28

!info Michel has a talk about using Enterprise Linux as a workstation - we're using Hyperscale as an incubator for GNOME backports until it takes off enough to get re-homed like the CPU SIG

<@salimma:fedora.im>

15:45:53

!info Davide and Michel has a community pavillion talk about how Meta uses CentOS Stream, Hyperscale will feature there too

<@salimma:fedora.im>

15:46:19

I don't have time to fish out links, since we only have 12 mins left

<@salimma:fedora.im>

15:46:27

!topic Tickets

<@salimma:fedora.im>

15:46:44

anything happened recently? let's take a quick look

<@ngompa:fedora.im>

15:46:55

we have a couple of kernel RFE tickets

<@ngompa:fedora.im>

15:47:06

they're waiting on RHEL Jiras being filed first before processing

<@salimma:fedora.im>

15:47:33

ps matrix.org has a status update now https://status.matrix.org/incidents/fhr7ky4ttlms

<@salimma:fedora.im>

15:47:36

they're still investigating

<@ngompa:fedora.im>

15:47:46

good to know

<@salimma:fedora.im>

15:49:18

there's the issue about documenting the GitLab workflows too

<@salimma:fedora.im>

15:49:40

but so far it seems that it works reasonably well even winging it

<@ngompa:fedora.im>

15:50:07

I'm hoping to get that up soon actually

<@salimma:fedora.im>

15:50:14

already better than the split situation of using Pagure but rebasing on top of an external repo (either Fedora's dist-git or GitLab) and not having working PRs

<@ngompa:fedora.im>

15:50:19

I need to figure out how much depth to document repo setup

<@salimma:fedora.im>

15:50:42

oh looks like... there's still the old feature request of enabling 9pfs, I can't remember why I filed it. let's see

<@salimma:fedora.im>

15:51:17

blocked on me filing JIRA. which I will try doing this week (still need to file a follow up asking why they didn't patch a second Emacs CVE)

<@salimma:fedora.im>

15:51:26

let's move on since the other two tickets I see are... membership

<@salimma:fedora.im>

15:51:31

!topic Membership

<@salimma:fedora.im>

15:51:38

!link https://pagure.io/centos-sig-hyperscale/sig/issue/185

<@salimma:fedora.im>

15:51:49

Vishal Mishra: want to introduce yourself?

<@vishalmishra434:fedora.im>

15:52:52

Hey Everyone,

<@vishalmishra434:fedora.im>

15:52:52

I am Vishal, I work at Meta and recently started focusing on openssh builds here. Excited to be here and looking forward to working with you all.

<@ngompa:fedora.im>

15:54:05

awesome!

<@salimma:fedora.im>

15:54:08

thanks Vishal

<@ngompa:fedora.im>

15:54:10

great to meet you

<@salimma:fedora.im>

15:54:35

I forgot, do we normally vote or we just ask for objections

<@ngompa:fedora.im>

15:54:50

the latter

<@salimma:fedora.im>

15:55:03

any objection?

<@ngompa:fedora.im>

15:55:08

none from me

<@salimma:fedora.im>

15:55:50

let's give it a minute, then we'll approve it and move on. There's another membership request but we don't have Tommy today so that will wait

<@salimma:fedora.im>

15:56:36

!info approved: welcoming Vishal to the Hyperscale SIG

<@salimma:fedora.im>

15:56:51

Neal Gompa (Fedora): want to process the ticket since I think you have access to the appropriate ACLs?

<@ngompa:fedora.im>

15:57:06

sure

<@ngompa:fedora.im>

15:57:12

and if I don't, I'll just bug Davide Cavalca

<@salimma:fedora.im>

15:57:26

yeah, I think he's not available today though

<@salimma:fedora.im>

15:57:31

!topic Miscellaneous

<@salimma:fedora.im>

15:57:39

I wonder if we should call this open floor like the other meetings

<@ngompa:fedora.im>

15:57:46

we probably should

<@ngompa:fedora.im>

15:57:53

also typing that word is hard 😅

<@salimma:fedora.im>

15:58:02

yeah that's why I was thinking of that :)

<@salimma:fedora.im>

15:58:22

I have one but it's more a CPU SIG thing really

<@salimma:fedora.im>

15:58:36

(which does not have its meeting yet - oh actually, the other part is probably relevant)

<@salimma:fedora.im>

15:59:13

1. I floated it at the last Hyperscale hangout, but we are considering making that a combined Hyperscale / CPU SIG hangout since the membership overlaps anyway and scheduling an extra event is difficult

<@salimma:fedora.im>

16:00:02

2. one thing CPU SIG might need to do differently than Hyperscale is... uh... if the package being updated is also released for i686 (ouch) on c9s / c10s - should it also release i686? how do we easily filter which package gets built for i686 and which one does not

<@ngompa:fedora.im>

16:00:21

we don't bother filtering

<@salimma:fedora.im>

16:00:28

# 2 is not theoretical, I just noticed freetype is actually available as i686 on c9s

<@salimma:fedora.im>

16:00:33

argh sorry, markdown

<@ngompa:fedora.im>

16:00:55

we will need CBS support for c9s i686

<@ngompa:fedora.im>

16:00:59

it currently does not exist

<@salimma:fedora.im>

16:01:11

in cases where it is installed, basically now you have two choices - either you can't use CPU SIG or you ... uninstall the i686

<@salimma:fedora.im>

16:01:19

ah ok, so we can just document this as a known limitation

<@salimma:fedora.im>

16:01:28

and... anyone else? I notice we're out of time

<@ngompa:fedora.im>

16:01:52

I've got nothing

<@ngompa:fedora.im>

16:01:58

thanks for running the meeting

<@ngompa:fedora.im>

16:02:03

(despite all the hiccups)

<@salimma:fedora.im>

16:02:54

np, not anyone's fault :)

<@salimma:fedora.im>

16:03:02

thanks all for coming!

<@salimma:fedora.im>

16:03:06

!endmeeting