<@amoloney:fedora.im>
17:01:54
!startmeeting F40 Beta Go/No-Go meeting
<@meetbot:fedora.im>
17:01:55
Meeting started at 2024-03-14 17:01:54 UTC
<@meetbot:fedora.im>
17:01:55
The Meeting name is 'F40 Beta Go/No-Go meeting'
<@amoloney:fedora.im>
17:02:28
!meetingname F40 Beta Go/No-Go-meeting
<@nirik:matrix.scrye.com>
17:02:38
morning
<@amoloney:fedora.im>
17:02:45
!info roll call
<@amoloney:fedora.im>
17:02:55
!hi
<@zodbot:fedora.im>
17:02:57
Aoife Moloney (amoloney)
<@jskladan:fedora.im>
17:03:03
!hi
<@zodbot:fedora.im>
17:03:04
Josef Skladanka (jskladan)
<@sgallagh:fedora.im>
17:03:54
!hi
<@zodbot:fedora.im>
17:03:56
Stephen Gallagher (sgallagh) - he / him / his
<@pwhalen:fedora.im>
17:05:44
!hi
<@zodbot:fedora.im>
17:05:46
Paul Whalen (pwhalen)
<@amoloney:fedora.im>
17:06:11
Ill give it another minute or two for folks to find their way here
<@coremodule:fedora.im>
17:06:14
!hi
<@zodbot:fedora.im>
17:06:17
Geoffrey Marr (coremodule)
<@adamwill:fedora.im>
17:08:04
!hi
<@zodbot:fedora.im>
17:08:05
Adam Williamson (adamwill) - he / him / his
<@amoloney:fedora.im>
17:08:46
!info purpose of this meeting
<@amoloney:fedora.im>
17:08:54
(boiler plate time)
<@amoloney:fedora.im>
17:09:07
!info Purpose of this meeting is to check whether or not F40 Beta is ready for shipment, according to the release criteria.
<@adamwill:fedora.im>
17:09:12
brb, switching room
<@amoloney:fedora.im>
17:09:30
!info this is determined in a few ways
<@bittin:fedora.im>
17:09:36
so are we go or no go? (confused by the US DST)
<@bittin:fedora.im>
17:09:47
!hi
<@sgallagh:fedora.im>
17:09:47
Luna Jernberg: The meeting just started
<@zodbot:fedora.im>
17:09:48
Luna Jernberg (bittin) - they / them / theirs
<@bittin:fedora.im>
17:09:55
Stephen Gallagher: ah alright i see :)
<@amoloney:fedora.im>
17:10:14
just going through the intro info right now :)
<@amoloney:fedora.im>
17:10:17
!info Release candidate compose is available
<@frantisekz:fedora.im>
17:10:19
!hi
<@zodbot:fedora.im>
17:10:21
František Zatloukal (frantisekz)
<@amoloney:fedora.im>
17:10:35
!info No remaining blocker bugs
<@amoloney:fedora.im>
17:10:41
!info Test matrices are complete
<@nirik:matrix.scrye.com>
17:10:44
(but spoiler, very likely no go. ;)
<@amoloney:fedora.im>
17:10:59
!info Fedora CoreOS and IoT are ready
<@amoloney:fedora.im>
17:11:09
!info Current Status: RC
<@nhanlon:beeper.com>
17:11:09
!hi
<@zodbot:fedora.im>
17:11:12
Neil Hanlon (neil) - he / him / his
<@amoloney:fedora.im>
17:11:25
Do we have an rc to discuss?
<@adamwill:fedora.im>
17:12:15
we do, i guess
<@adamwill:fedora.im>
17:12:16
we have a lot of them!
<@lruzicka:matrix.org>
17:12:49
Allo
<@sgallagh:fedora.im>
17:13:22
adamw: Water, water everywhere...
<@amoloney:fedora.im>
17:14:43
any of them good enough to be a Beta release?
<@amoloney:fedora.im>
17:14:50
* keeps fingers crossed
<@adamwill:fedora.im>
17:14:56
well, i think the other topics cover that
<@adamwill:fedora.im>
17:15:12
we put this one up front just on the logic that "if we have no rc we're clearly no-go"
<@amoloney:fedora.im>
17:15:14
so shall I move on to blodkers?
<@adamwill:fedora.im>
17:15:15
if we *do* have an rc, we move on
<@adamwill:fedora.im>
17:15:20
yup
<@amoloney:fedora.im>
17:15:24
ah cool thank you :)
<@amoloney:fedora.im>
17:15:46
!info we do have an RC for F40 Beta, moving on to Blocker Status
<@amoloney:fedora.im>
17:15:59
!info Blockers
<@amoloney:fedora.im>
17:16:30
!info link https://qa.fedoraproject.org/blockerbugs/milestone/40/beta/buglist
<@amoloney:fedora.im>
17:16:42
!chair adamw
<@amoloney:fedora.im>
17:16:53
darn I dont know that one works
<@amoloney:fedora.im>
17:17:01
anyway turning this over to you Adam
<@nhanlon:beeper.com>
17:17:30
chair is not implemented in the current meetbot, so no worries
<@sgallagh:fedora.im>
17:17:43
Right, anyone can use the meetbot commands
<@amoloney:fedora.im>
17:17:53
another learning :)
<@adamwill:fedora.im>
17:17:54
gkimm e a sec
<@nhanlon:beeper.com>
17:18:00
adam told me I wasn't allowed...
<@sgallagh:fedora.im>
17:18:20
Anyone can use the meetbot commands from a *technical* perspective.
<@amoloney:fedora.im>
17:18:27
the command to disallow Neil Hanlon *is* enabled :p
<@adamwill:fedora.im>
17:20:27
we do have a couple of proposed blockers
<@adamwill:fedora.im>
17:20:33
!topic (2269407) Fedora minimal does not boot on Raspberry Pi 4b+
<@adamwill:fedora.im>
17:20:37
<@adamwill:fedora.im>
17:20:41
<@adamwill:fedora.im>
17:20:43
!info Proposed Blocker, bcm283x-firmware, NEW
<@adamwill:fedora.im>
17:20:55
this should probably be titled "osbuild-built minimal image does not boot on (something)"
<@adamwill:fedora.im>
17:21:00
not sure if we actually tried the image on other hw yet
<@adamwill:fedora.im>
17:21:46
it's pretty blockery, the fix would be to fall back to the older image build tool
<@adamwill:fedora.im>
17:21:51
Peter Robinson: around?
<@adamwill:fedora.im>
17:22:16
Josef Skladanka: do you know if we tested on any other hw yet?
<@lruzicka:matrix.org>
17:22:46
I dont know
<@jskladan:fedora.im>
17:22:47
Don't know. František Zatloukal ?
<@adamwill:fedora.im>
17:23:57
the image *does* boot on a VM, btw
<@frantisekz:fedora.im>
17:24:12
can reproduce that on our rpi4 too
<@pwhalen:fedora.im>
17:24:16
I've not tested other hardware, trying to work out the problem with it
<@frantisekz:fedora.im>
17:24:26
osbuild minimal image doesn't boot up
<@adamwill:fedora.im>
17:24:34
but anyhoo...for a vote, I'm +1
<@frantisekz:fedora.im>
17:24:48
+1
<@nirik:matrix.scrye.com>
17:24:51
+1 here...
<@nhanlon:beeper.com>
17:24:59
I can test this on some of my PIs. but, +1
<@adamwill:fedora.im>
17:25:18
Neil Hanlon: i'd be kinda interested to see the result on, like, a jetson nano or something
<@adamwill:fedora.im>
17:25:24
sadly i'm travelling so can't try on mine
<@pwhalen:fedora.im>
17:25:49
adamwI'll try as well
<@adamwill:fedora.im>
17:25:56
thanks
<@nhanlon:beeper.com>
17:25:57
I have a handful of non-pi boards I can try. sadly not a jetson nano though
<@adamwill:fedora.im>
17:26:04
anything non-pi is good i guess
<@adamwill:fedora.im>
17:26:25
just kinda interested if this is some kind of 'everything outside of vm-style UEFI boot is broken' or it's somehow more pi-specific
<@nhanlon:beeper.com>
17:26:39
I can test Pine64, RockPro64, Libre Computer boards (4x)
<@nhanlon:beeper.com>
17:27:01
will ask rocky altarch sig to test too, they have orange pi, banana pi
<@adamwill:fedora.im>
17:27:32
proposed #agreed 2269407 - AcceptedBlocker (Beta) - this is accepted as a violation of "Release-blocking ARM disk images must boot to the initial-setup utility" on the Pi 4b+. Minimal is a release-blocking image, and Pi 4s are release-blocking ARM devices
<@nhanlon:beeper.com>
17:27:52
is there a link to the latest image?
<@adamwill:fedora.im>
17:27:55
ugh
<@adamwill:fedora.im>
17:27:58
proposed !agreed
<@adamwill:fedora.im>
17:27:59
:P
<@geraldosimiao:matrix.org>
17:28:14
!hi
<@nirik:matrix.scrye.com>
17:28:17
ack (aside the octothorpe issue)
<@zodbot:fedora.im>
17:28:17
Geraldo S. Simião Kutz (geraldosimiao) - he / him / his
<@adamwill:fedora.im>
17:28:21
Neil Hanlon: https://kojipkgs.fedoraproject.org/compose/40/Fedora-40-20240313.1/compose/Spins/aarch64/images/Fedora-Minimal-40_Beta-1.4.aarch64.raw.xz
<@nhanlon:beeper.com>
17:28:22
ack.
<@nhanlon:beeper.com>
17:28:29
ty!
<@lruzicka:matrix.org>
17:28:32
ack
<@geraldosimiao:matrix.org>
17:28:32
Ack
<@nirik:matrix.scrye.com>
17:28:40
is that the osbuild one?
<@jskladan:fedora.im>
17:28:41
ack
<@adamwill:fedora.im>
17:28:47
yes
<@adamwill:fedora.im>
17:28:52
the last non-osbuild was the 12.n.0 nightly iirc
<@adamwill:fedora.im>
17:28:58
so compare to that for a sanity check
<@nirik:matrix.scrye.com>
17:29:04
yeah.
<@adamwill:fedora.im>
17:31:59
oh sorry
<@adamwill:fedora.im>
17:32:05
i was waiting for someone else to do something for some reason :D
<@adamwill:fedora.im>
17:32:11
!agreed 2269407 - AcceptedBlocker (Beta) - this is accepted as a violation of "Release-blocking ARM disk images must boot to the initial-setup utility" on the Pi 4b+. Minimal is a release-blocking image, and Pi 4s are release-blocking ARM devices
<@nhanlon:beeper.com>
17:32:15
😂
<@adamwill:fedora.im>
17:32:30
!topic (2269373) Fedora Media Writer (Windows Version) writes images that fail the checksum check on boot
<@adamwill:fedora.im>
17:32:33
<@adamwill:fedora.im>
17:32:37
<@adamwill:fedora.im>
17:32:40
!info Proposed Blocker, mediawriter, NEW
<@adamwill:fedora.im>
17:32:43
!info Ticket vote: BetaBlocker (+0,0,-3) (-geraldosimiao, -kparal, -sgallagh)
<@adamwill:fedora.im>
17:32:52
seems like we kinda have a consensus that we don't really need to track this as a blocker
<@geraldosimiao:matrix.org>
17:33:10
Yeah
<@sgallagh:fedora.im>
17:33:11
Yeah, I proposed it as such when I initially thought that the written images were actually broken.
<@adamwill:fedora.im>
17:33:12
anyone want to argue for blocker?
<@lruzicka:matrix.org>
17:33:22
Nope, windows application :D
<@sgallagh:fedora.im>
17:34:01
I do kind of want to suggest that maybe the default boot should *not* be the verification mode, since this is definitely going to keep happening to folks trying it out from Windows.
<@sgallagh:fedora.im>
17:34:13
But we can discuss that elsewhere
<@lruzicka:matrix.org>
17:34:14
If, maybe a final one
<@adamwill:fedora.im>
17:34:22
we've kicked that around before
<@adamwill:fedora.im>
17:34:48
proposed !agreed 2269373 - RejectedBlocker (Beta) - this is rejected as not violating any release criteria
<@sgallagh:fedora.im>
17:34:56
ack
<@geraldosimiao:matrix.org>
17:35:05
ack
<@lruzicka:matrix.org>
17:35:06
ack
<@jskladan:fedora.im>
17:35:25
ack
<@adamwill:fedora.im>
17:35:27
!agreed 2269373 - RejectedBlocker (Beta) - this is rejected as not violating any release criteria
<@adamwill:fedora.im>
17:36:05
and we should note that we have four other outstanding, unaddressed blockers:
<@adamwill:fedora.im>
17:36:13
!info other outstanding, unaddressed blockers:
<@adamwill:fedora.im>
17:36:34
!info (2267968) Raspberry Pi 400 shows nothing on screen when booting Fedora 40 images (even before grub)
<@adamwill:fedora.im>
17:36:38
<@adamwill:fedora.im>
17:36:48
!info (2269385) rhgb breaks custom/minimal install on most filesystem layouts
<@adamwill:fedora.im>
17:36:52
<@adamwill:fedora.im>
17:37:00
!info (2113005) Live image made with BOOTX64.EFI from latest shim-x64-15.6-2 fails to boot on some boards
<@adamwill:fedora.im>
17:37:06
<@adamwill:fedora.im>
17:37:18
!info (2259264) Fedora fails to boot via BOOT/bootaa64->fbaa64 on UEFI machines with EFI_MEMORY_ATTRIBUTES_PROTOCOL
<@adamwill:fedora.im>
17:37:21
<@adamwill:fedora.im>
17:37:41
so...unless we want to consider waiving every single one of those plus the arm minimal one...we can probably go to a decision at this point?
<@geraldosimiao:matrix.org>
17:37:59
A blocker frenzy
<@adamwill:fedora.im>
17:38:03
does anyone want to waive all of those :D
<@amoloney:fedora.im>
17:38:11
a....bloc-party?
<@amoloney:fedora.im>
17:38:17
sorry :D
<@amoloney:fedora.im>
17:38:31
but to answer Adams q, no
<@lruzicka:matrix.org>
17:38:40
I think this should be a matter of decision of how much we really need to block the release for ARM.
<@adamwill:fedora.im>
17:38:51
well, the rhgb bug is not arm
<@nirik:matrix.scrye.com>
17:38:57
There has been a discussion about that...
<@adamwill:fedora.im>
17:38:58
but yes, the other non-shim bugs are all arm
<@sgallagh:fedora.im>
17:39:00
lruzicka: As of right now, ARM remains a blocking arch
<@frantisekz:fedora.im>
17:39:13
that rhgb thing is late blocker though! 😅
<@geraldosimiao:matrix.org>
17:39:32
The rhgb bug meets the blocking criteria
<@adamwill:fedora.im>
17:39:50
it's a waiver candidate, but i would vote against waiving it i think
<@lruzicka:matrix.org>
17:39:55
I know, but I guess we would not have to hold this meeting if people were not fiddling with the thought of waiving.
<@sgallagh:fedora.im>
17:40:26
The rhgb issue is enough to block the release, to my mind.
<@sgallagh:fedora.im>
17:40:36
Even coming in late, it's too big to ignore
<@geraldosimiao:matrix.org>
17:40:45
It seems a week more on this can help
<@nirik:matrix.scrye.com>
17:40:52
I guess even if we waived everything, really the only thing we could ship is 1.2 right? (the non kiwi build one)?
<@adamwill:fedora.im>
17:41:19
why not the kiwi build ones again? i am forgetting stuff
<@adamwill:fedora.im>
17:41:24
but...it seems kinda academic
<@nirik:matrix.scrye.com>
17:42:22
the only one where they exist is 1.5 and it's DOOMED due to toolbox failing.
<@nirik:matrix.scrye.com>
17:42:45
and we can't actually ship 1.2 anyhow, because we added some packages to fix some blockers after it.
<@nirik:matrix.scrye.com>
17:42:49
so, yeah... no go.
<@geraldosimiao:matrix.org>
17:43:40
Yeah, I agree. Can't ship beta like this.
<@amoloney:fedora.im>
17:45:57
shall I go to the formal vote then?
<@nirik:matrix.scrye.com>
17:46:46
sure.
<@sgallagh:fedora.im>
17:47:13
I vote Go! ... to starting the formal vote ;-)
<@amoloney:fedora.im>
17:47:51
!topic F40 Beta Go/No-Go decision
<@amoloney:fedora.im>
17:48:09
I will poll each team. Please reply 'Go' or 'No Go'
<@amoloney:fedora.im>
17:48:15
FESCo?
<@sgallagh:fedora.im>
17:48:18
No Go
<@nirik:matrix.scrye.com>
17:48:20
No Go
<@amoloney:fedora.im>
17:48:40
Releng?
<@jnsamyak:matrix.org>
17:49:02
No Go
<@jnsamyak:matrix.org>
17:49:07
:D
<@amoloney:fedora.im>
17:49:41
QA?
<@lruzicka:matrix.org>
17:49:52
No go
<@jskladan:fedora.im>
17:49:56
No Go
<@geraldosimiao:matrix.org>
17:49:57
No Go
<@amoloney:fedora.im>
17:50:22
!agreed The Fedora Linux 40 Beta is NO-GO
<@amoloney:fedora.im>
17:50:59
!info The next F40 Beta Go/No-Go meeting will be on Thursday 21st March @ 1700 UTC
<@nirik:matrix.scrye.com>
17:51:13
Hopefully we can be in better shape then...
<@jnsamyak:matrix.org>
17:51:31
🤞
<@amoloney:fedora.im>
17:51:34
!info F40 Beta will now shift to target date #2: Tuesday 2024-03-26
<@adamwill:fedora.im>
17:51:37
yep
<@amoloney:fedora.im>
17:52:05
!action amoloney to announce decision and adjust schedule
<@amoloney:fedora.im>
17:52:30
I wonder should I have @'d myself there...but anyway, I know I have to do it :)
<@amoloney:fedora.im>
17:52:44
and now...
<@amoloney:fedora.im>
17:52:50
!topic Open Floor
<@amoloney:fedora.im>
17:53:04
anything else to discuss before we wrap?
<@adamwill:fedora.im>
17:54:40
not...really, i guess just this has been a great lesson in why we shouldn't try and change two major image build systems during freeze :P
<@lruzicka:matrix.org>
17:54:51
I think we should consider a change to the release cycle that keeps everything frozen from Beta freeze until Final so that we can really focus on nuances and stuff.
<@lruzicka:matrix.org>
17:55:12
Otherwise, we will be always finding blockers late in the process.
<@amoloney:fedora.im>
17:55:50
would there be anything to be said about doing a little retrospective after this release?
<@amoloney:fedora.im>
17:56:13
I dont know how yet....but something like that might be useful...
<@nirik:matrix.scrye.com>
17:56:42
yeah, probibly we shouldn't have tried to land those things, but oh well, we did...
<@nirik:matrix.scrye.com>
17:57:02
retro is good.
<@nirik:matrix.scrye.com>
17:57:26
and if we want to adjust the release cycle we can, but I don't think a extended freeze will go over too well.
<@adamwill:fedora.im>
17:57:32
looks like someone worked out the osbuild arm bug, yay
<@nhanlon:beeper.com>
17:57:44
there might be riots :D
<@adamwill:fedora.im>
17:58:04
lruzicka: i don't think that's practical. maintainers would revolt. people already hate the existing freezes
<@adamwill:fedora.im>
17:58:24
Aoife Moloney: we used to do retrospectives in qa, kinda dropped the ball on those lately
<@frantisekz:fedora.im>
17:58:44
what would be the harder way to handle that and not annoy maintainers - freeze only stuff affecting blocking deliverables...
<@nirik:matrix.scrye.com>
17:58:46
it's also possible with CI/gating we could do less freezing...
<@lruzicka:matrix.org>
17:58:56
I know. But this somehow feels like trying to extinguish a forest fire. When you think you are done, it starts elsewhere.
<@adamwill:fedora.im>
17:59:29
lruzicka: i'm not sure i agree? i don't really recall many cases where we had a flood of blocker bugs introduced between beta and final? maybe i'm not forgetting them
<@adamwill:fedora.im>
17:59:47
i mean, it *happens*, but it doesn't stick out in my mind as a huge problem
<@lruzicka:matrix.org>
18:00:38
Well, I remember something like a couple releases ago when we were having like three go/no goes meetings for Final and we always found another blocker on Thursday, that was quite frustrating.
<@amoloney:fedora.im>
18:00:41
things phase in and out of being needed, it might be time to look at doing some kind of retro again and more holistically across the release teams ( I know thats not a term, but I dont know what else to call everyone ) :)
<@geraldosimiao:matrix.org>
18:00:47
Some last hour blocker bugs aren't with brand new packages. Usually is just something we dind saw earlier
<@geraldosimiao:matrix.org>
18:01:40
Some last hour blocker bugs aren't with brand new packages. Usually is just something we didn't saw earlier
<@geraldosimiao:matrix.org>
18:02:00
Some last hour blocker bugs aren't with brand new packages. Usually is just something we didn't catch earlier
<@amoloney:fedora.im>
18:02:31
!idea #proposed: try to do a retrospective of the F40 release
<@nirik:matrix.scrye.com>
18:02:58
Having a longer time with a rc avialable helps... then it can get a lot more testing of corner cases.
<@lruzicka:matrix.org>
18:03:18
We usually get one day. Two at most.
<@adamwill:fedora.im>
18:03:22
i know i say this every cycle but we really need to test stuff *before* candidates
<@adamwill:fedora.im>
18:03:29
we should've caught the rhgb thing earlier
<@adamwill:fedora.im>
18:03:38
(maybe i would've noticed it if i wasn't spending all my time fixing kiwi stuff, sigh
<@adamwill:fedora.im>
18:04:15
but if we just can't get testing of nightlies, maybe we'll have to do more non-rc candidates
<@lruzicka:matrix.org>
18:04:40
I caught it on PiKVM when testing something else.
<@frantisekz:fedora.im>
18:05:17
maybe, it'd make sense to start doing rcs even with known blockers? simply once the freeze starts? it may attract more testing... idk
<@nirik:matrix.scrye.com>
18:05:38
at that point... they would be nightlys right?
<@nirik:matrix.scrye.com>
18:06:06
Perhaps a call for testing nightly right after freeze starts?
<@adamwill:fedora.im>
18:06:40
František Zatloukal: that's what a 'non-rc candidate' is
<@adamwill:fedora.im>
18:06:46
i don't like doing them because they are kind of pointless
<@adamwill:fedora.im>
18:06:49
theoretically speaking
<@adamwill:fedora.im>
18:06:54
but if it would make people test, sure, we can do them
<@nirik:matrix.scrye.com>
18:07:18
cp -a development/40 alt/stage/SUPER_GREAT_RC1.0
<@adamwill:fedora.im>
18:07:26
nirik: it would at least let releng shake the bugs out of the candidate script a bit earlier i guess :P
<@nhanlon:beeper.com>
18:07:31
it might be interesting to see if people don't feel a branch is "ready" until it's a "Candidate"
<@adamwill:fedora.im>
18:07:33
nirik: hehe
<@nirik:matrix.scrye.com>
18:08:02
I was thinking about that. If we could ever get time, it would be nice to get staging working for composes... so we could test there.
<@geraldosimiao:matrix.org>
18:08:35
Names, names, what the right nomenclature can do for us...
<@geraldosimiao:matrix.org>
18:09:06
But yeah, I agree
<@amoloney:fedora.im>
18:11:38
Is there a summary I should/could take from this discussion? Or is it more just general 'musings among friends for a better release world'? :)
<@nirik:matrix.scrye.com>
18:11:56
I think discussion should go do list or... discussion. ;)
<@geraldosimiao:matrix.org>
18:12:46
Discussion is fine
<@adamwill:fedora.im>
18:13:16
it was musing
<@adamwill:fedora.im>
18:13:17
we like to muse
<@lruzicka:matrix.org>
18:13:40
But sometimes it would be nice if the world became a better place.
<@nhanlon:beeper.com>
18:13:49
i like Muse
<@amoloney:fedora.im>
18:14:11
That world is in a place, somewhere only we know...
<@geraldosimiao:matrix.org>
18:14:59
😁
<@lruzicka:matrix.org>
18:15:48
Too complicated for me, Musescore, Lilypond, Ardour ...
<@amoloney:fedora.im>
18:17:08
ok on that *note* I think this meeting is at its end
<@amoloney:fedora.im>
18:17:11
!endmeeting