#meeting:fedoraproject.org log

<@davide:cavalca.name>

15:09:16

!startmeeting CentOS Hyperscale SIG

<@meetbot:fedora.im>

15:09:18

Meeting started at 2025-03-12 15:09:16 UTC

<@meetbot:fedora.im>

15:09:18

The Meeting name is 'CentOS Hyperscale SIG'

<@davide:cavalca.name>

15:09:46

morning folks

<@davide:cavalca.name>

15:09:52

apologies for the delay, DST is hard

<@davide:cavalca.name>

15:09:53

!topic Roll call

<@salimma:fedora.im>

15:10:05

!hi

<@zodbot:fedora.im>

15:10:06

Michel Lind (salimma) - he / him / his

<@rcolebaugh:matrix.org>

15:10:14

!hi

<@zodbot:fedora.im>

15:10:17

Raymond Colebaugh (rcolebaugh) - he / him / his

<@salimma:fedora.im>

15:10:45

DST is hard enough for me here in central time, must be rough having an 8am meeting that feels like 7am on the West Coast

<@davide:cavalca.name>

15:13:24

let's get started

<@davide:cavalca.name>

15:13:24

!topic Followups

<@davide:cavalca.name>

15:13:34

anybody has followups to share from the last meeting?

<@salimma:fedora.im>

15:14:22

follow up re: calendering

<@salimma:fedora.im>

15:14:47

the Fedora one seems to work fine, but I forgot to get us removed from the CentOS calendar so I'll do that after this

<@salimma:fedora.im>

15:15:10

(for those not in the know, that one insists everything is UTC and needs someone to manually merge a PR to update so it's a PITA)

<@davide:cavalca.name>

15:17:21

afaik shaunm was planning to sunset the CentOS calendar entirely in favor of just using the Fedora one

<@salimma:fedora.im>

15:18:26

yeah, I think we're the guinea pig I guess

<@salimma:fedora.im>

15:19:01

I can submit the PR then it's up to him to merge or not - but in the meantime if you don't need to track another CentOS SIG I'd suggest unsubscribing from that calendar

<@salimma:fedora.im>

15:19:32

if nobody else has something I have a second follow up

<@salimma:fedora.im>

15:20:22

!info Hyperscale SIG meetings tracked in calendar.fedoraproject.org, ignore the entries in the CentOS calendar

<@salimma:fedora.im>

15:21:31

the Emacs 30.1 build seems to work fine, we're still rolling it out slowly at work due to an unrelated issue. we beat Fedora stable releases (to be fair they needed FESCo approval to update stable releases)

<@salimma:fedora.im>

15:22:07

!info Emacs 30.1 works fine, including the libtree-sitter soname bump

<@davide:cavalca.name>

15:23:52

thanks

<@davide:cavalca.name>

15:23:57

next up

<@davide:cavalca.name>

15:23:58

!topic Announcements

<@davide:cavalca.name>

15:24:04

a bunch of us went to SCALE, it was fun

<@davide:cavalca.name>

15:24:18

and some of us will be at Red Hat Summit later in May

<@davide:cavalca.name>

15:24:31

anybody else has announcements to share?

<@daandemeyer/:matrix.org>

15:25:17

Ivan did a bunch of work on CI/CD in the gitlab systemd repo

<@daandemeyer/:matrix.org>

15:25:55

We're running the upstream integration tests on systemd builds from CBS now n every MR

<@daandemeyer/:matrix.org>

15:27:41

Biggest pain point is probably still testing-farm. While it works, life would be much easier if we could use the gitlab hosted runners, but AFAIK we don't have any avenue to get gitlab to implement a one line change for us to enable nested virtualization on their hosted runners

<@davide:cavalca.name>

15:30:35

do you know if that's a gitlab feature change, or just a config that's specific to our deployment?

<@daandemeyer/:matrix.org>

15:30:45

Feature change unfortunately

<@daandemeyer/:matrix.org>

15:31:00

https://gitlab.com/gitlab-org/gitlab-runner/-/issues/37193

<@salimma:fedora.im>

15:31:17

speaking of which - we still don't have it formalized yet, but I've started using gitlab for packages recently too and it's a much nicer workflow than pagure (partly because the Stream and HS repos are all in one forge now, and we can do mirroring)

<@salimma:fedora.im>

15:31:37

also looking forward to being able to /merge/ PRs instead of having to just close them and merge by hand ;)

<@salimma:fedora.im>

15:32:39

ouch... one year without any update?

<@daandemeyer/:matrix.org>

15:33:23

The biggest problem this would solve is that we could enable PRS from forks from non-sig members since we wouldn't need the testing farm token to be available in MR jobs and could build rpms locally for running the tests instead of using cbs

<@daandemeyer/:matrix.org>

15:33:59

Which would mean that anyone outside of the SIG could trivially contribute, instead of having to be a sig member

<@daandemeyer/:matrix.org>

15:34:24

Currently people outside the sig can make PRs, but the tests won't run because we can't give people outside the sig access to the secrets

<@salimma:fedora.im>

15:35:10

RH hasn't solved this for people trying to contribute to Stream too right

<@davide:cavalca.name>

15:35:19

I will try and find out again who owns the gitlab commercial relationship and see if we can apply pressure that way

<@salimma:fedora.im>

15:35:33

on the RH side?

<@davide:cavalca.name>

15:35:56

and on the CentOS side (as CentOS is technically under a different contract)

<@salimma:fedora.im>

15:37:08

ahh

<@daandemeyer/:matrix.org>

15:37:37

With the gitlab ultimate contract we probably have a ridiculous amount of hardware resources at our disposal in gitlab

<@daandemeyer/:matrix.org>

15:37:41

We just can't make use of them atm which is a bit sad

<@salimma:fedora.im>

15:38:01

I suspect RHEL has a different contract than CentOS, right

<@salimma:fedora.im>

15:38:14

or maybe it's the same since they need to run the same gating tests pretty much

<@daandemeyer/:matrix.org>

15:38:19

This only applies to systemd ofc, for all the other stuff you don't need nested virt

<@daandemeyer/:matrix.org>

15:38:44

(well kernel could as well, but generally this isn't needed)

<@salimma:fedora.im>

15:43:00

we've been quiet for a while - any other announcement?

<@davide:cavalca.name>

15:43:13

nothing else from me

<@salimma:fedora.im>

15:43:37

There's a Stream office hour after this, I think

<@davide:cavalca.name>

15:43:56

oh I don't have that on my calendar

<@salimma:fedora.im>

15:43:58

https://github.com/CentOS/Calendar/blob/main/meetings/stream-office-hours.yaml

<@davide:cavalca.name>

15:44:32

anyway, next up

<@davide:cavalca.name>

15:44:37

!topic Tickets

<@davide:cavalca.name>

15:44:51

we still need to clean up the tickets :)

<@salimma:fedora.im>

15:48:01

yeah

<@salimma:fedora.im>

15:48:14

we should chomp through them at the next hangout if we have not gone through them by then?

<@rcolebaugh:matrix.org>

15:48:51

Nice that sounds good to me

<@salimma:fedora.im>

15:51:42

we can move on I think

<@salimma:fedora.im>

16:02:46

So Davide said his Matrix server died

<@salimma:fedora.im>

16:03:02

!topic Membership

<@salimma:fedora.im>

16:03:12

@rcolebaugh is Vishal here?

<@jonathanspw:fedora.im>

16:03:29

!hi

<@zodbot:fedora.im>

16:03:30

Jonathan Wright (jonathanspw)

<@rcolebaugh:matrix.org>

16:04:22

No I didn't think so, hopefully he can make it next time. He did file a membership ticket: https://pagure.io/centos-sig-hyperscale/sig/issue/185

<@salimma:fedora.im>

16:04:44

Yeah, if he doesn’t join this meeting we can’t really onboard him

<@salimma:fedora.im>

16:04:59

File a special one off calendar invite I guess?

<@rcolebaugh:matrix.org>

16:05:17

Sounds good, I can set that up!

<@salimma:fedora.im>

16:05:36

Let’s wrap this up since we’re over time. Sorry, technical snafu on my end too so I couldn’t check both matrix and the internal chat at the same time and we lost 10 mins

<@salimma:fedora.im>

16:05:42

!endmeeting