17:00:22 #startmeeting FESCO (2023-09-28) 17:00:22 Meeting started Thu Sep 28 17:00:22 2023 UTC. 17:00:22 This meeting is logged and archived in a public location. 17:00:22 The chair is sgallagh. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 17:00:22 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:22 The meeting name has been set to 'fesco_(2023-09-28)' 17:00:23 #meetingname fesco 17:00:23 The meeting name has been set to 'fesco' 17:00:23 #chair nirik, decathorpe, zbyszek, sgallagh, mhroncok, dcantrell, mhayden, Conan_Kudo, Pharaoh_Atem, Son_Goku, King_InuYasha, Sir_Gallantmon, Eighth_Doctor, tstellar 17:00:23 Current chairs: Conan_Kudo Eighth_Doctor King_InuYasha Pharaoh_Atem Sir_Gallantmon Son_Goku dcantrell decathorpe mhayden mhroncok nirik sgallagh tstellar zbyszek 17:00:24 #topic init process 17:00:24 .hi 17:00:27 .hello salimma 17:00:29 sgallagh: sgallagh 'Stephen Gallagher' 17:00:35 michel-slm: salimma 'Michel Lind' 17:00:37 .hi 17:00:42 decathorpe: decathorpe 'Fabio Valentini' 17:01:00 .hello ngompa 17:01:06 Son_Goku: ngompa 'Neal Gompa' 17:01:11 morning. 17:01:15 .hello kevin 17:01:21 nirik: kevin 'Kevin Fenzi' 17:01:27 .hello2 17:01:35 zbyszek: zbyszek 'Zbigniew Jędrzejewski-Szmek' 17:02:39 .hello tstellar 17:02:44 tstellar: tstellar 'Tom Stellard' 17:04:42 We have quorum, if just barely. 17:05:03 I'm going to wait a couple minutes more (I know mhayden was going to be a little late) 17:07:28 OK, let's get started. 17:07:32 #topic #3059 F39 incomplete changes: 100% complete deadline 17:07:33 .fesco 3059 17:07:34 sgallagh: Issue #3059: F39 incomplete changes: 100% complete deadline - fesco - Pagure.io - https://pagure.io/fesco/issue/3059 17:08:29 so where we here? 17:08:32 We had some action items from the previous week… 17:08:45 I'll be honest; I'm not sure what is still remaining here; without our PgM, this ticket hasn't been staying up to date. 17:08:49 #info https://src.fedoraproject.org/rpms/fedora-release/pull-request/279 has been merged. 17:09:05 .hello2 17:09:11 mhayden: mhayden 'Major Hayden' 17:09:17 yeah, thats rawhide only though right? 17:09:21 Oh, but it actually needs fixing. 17:09:27 https://src.fedoraproject.org/rpms/fedora-release/pull-request/279#comment-159648 17:09:30 we need a 39 cherry-pick or update 17:09:53 yes, and the fwupd fix 17:11:27 mhayden Did the Azure image publication get handled? 17:11:38 zbyszek: can you do that? I'm not sure I have more cycles right now... ;( 17:11:46 I don't see any changes in fwupd spec file (or pull requests). 17:11:49 sgallagh: davedunc was going to upload the betas 17:12:09 i also spoke to themayor this week via video to see if Azure could help us meet in the middle to make it easier 17:12:22 * mhayden scurries off to check the azure account 17:12:45 oh, I suppose fwupd is one of the secure-boot packages that needs someone in that group to do. 17:13:18 OK, I'll try to file a pull request. 17:13:50 ok, I can try and shepard out a build. Should I just cherry pick that commit for fedora-release/f39? 17:14:28 nirik: yeah 17:14:54 or sgallagh can since he's been doing fedora-release builds? ;0 17:15:11 hm ... is adamw's summary here still up-to-date, or is that from before last week's meeting? https://pagure.io/fesco/issue/3059#comment-875161 17:15:29 sgallagh: azure images aren't in place yet. i think we can get automation done for it in F40 and do it manually in the meantime 17:15:39 decathorpe: from before 17:15:49 Yeah, I'll take care of the F39 backport in fedora-release 17:16:01 Though I'm unclear on whether I should do that now or wait for zbyszek's PR? 17:16:30 now 17:16:39 the fwupd thing just does it for upgrades... 17:17:02 (well, will when it exists) 17:18:02 #action sgallagh to backport the fwupd patch to F39 in fedora-release 17:18:39 mhayden: Can you take an action to get the Beta in place manually by next meeting? 17:20:29 OK, I don't think there's anything further to discuss on this topic this week. On to New Business 17:20:40 #topic #3065 New workstation WebUI is enforcing a requirement for a separate /boot 17:20:41 .fesco 3065 17:20:42 sgallagh: Issue #3065: New workstation WebUI is enforcing a requirement for a separate /boot - fesco - Pagure.io - https://pagure.io/fesco/issue/3065 17:22:22 Dunno, maybe we should just let the discussion proceed in the bug? This will not be a big issue before F40. 17:22:33 zbyszek: You proposed it for the meeting :) 17:22:41 +1 17:22:42 Yeah, because we have a ticket open. 17:23:00 I'd say close the ticket and ask them to file new/reopen if there's still anything to decide before f40? 17:23:09 I mean: people ask us for input, we should answer. Even if the answer is that we don't take a stance. 17:23:19 nirik: +1 17:23:42 To be clear (and for the logs): F39 does not have a requirement on `/boot` to be its own partition, correct? 17:24:28 It doesn't. 17:24:29 "On traditional anaconda interface you can install fedora using a partition layout without a separate partition for the /boot mount point (using ext4 or even btrfs) and the installed system works normally, without problems." 17:25:05 zbyszek: agreed, seems like it'd be nice if fesco can either say this is fine, this is not fine, or it's up to the Anaconda devs 17:25:55 michel-slm: that's what I suggested in the ticket too ;) 17:26:15 But that was 15 days ago. 17:26:19 My two cents: It's desirable for a number of reasons to have it be separate. The default layout must include a separate /boot. The custom configuration must allow it to be excluded. 17:26:28 Fully loaded foot-guns notwithstanding 17:26:57 I don't think the anaconda team has really answered anyone in the bug yet have they? aside from the initial query as to what the use case was? 17:27:19 what sgallagh proposed sounds nice - does the Web UI allow that now? 17:27:41 From the comments on the ticket, it sounds like it does not 17:27:47 yeah 17:28:05 having to resort to kickstarts to configure this sounds like a regression 17:28:47 If you're installing via the graphical installer, do we really need to support such custom cases? 17:29:25 I think I'd be fine with the Anaconda team saying that this is too much hassle and you need to use a kickstart if you want to do this. It's mostly useful for very small images. 17:29:35 There's interface concerns here too... like the web installer doesn't have a 'recommend' type thing... 17:29:36 zbyszek From personal experience, I find that much of the time the first test deployments will be done via the GUI and then kickstarts will be modified from that. 17:29:54 So I'd say we *probably* don't want to relegate that to kickstart-only 17:30:07 well, not supporting it in the Web UI when it was supported in the GTK UI is a regression, isn't it? 17:30:44 * nirik tries to avoid getting on his soap box about regressions. I should probibly give up. 17:30:46 I'd treat it as such, whether there are valid use cases or not (which is always hard to determine). somebody *will* complain if it used to work but no longer odes :) 17:31:03 it's a regression 17:31:17 decathorpe: Removal of functionality is not always the same thing as a regression. Particularly if done deliberately and with consideration. 17:31:40 well, if it *is* deliberate, that needs to be communicated 17:31:58 No disagreement on that point 17:32:22 ok, I'll do it one more time. A regression is when: a) There is a published specification. b) everyone has agreed to abide by it. c) The product was tested and found to match the specification. d) some later test shows that it doesn't match the specification. 17:32:51 alright, alright, I regresst my case 17:33:14 nirik: I have thoughts, but they're a tangent so I will defer them 17:33:15 sadly, it's like hacker... so I should stop. ;) 17:33:35 so this is like the cpython API - or gnome extensions - since there was never a spec, it's never a regression :) 17:33:45 In https://bugzilla.redhat.com/show_bug.cgi?id=2234640#c1, Jiri says that " However, yes, we should allow to use BTRFS subvolume if that is not possible right now.". 17:33:56 anyhow, I personally would prefer anaconda developers and users work out things... if they cannot get to any agreement by before f40 beta, we can look and see if we want to ask them to do something 17:34:05 That's fair. 17:34:12 So I'm still +1 to nirik proposal. 17:34:16 Let me formalize that in a proposal... 17:35:53 Proposal: If Anaconda devs feel strongly about mandating a separate /boot partition/subvolume, FESCo would like to see them propose a Fedora 40 Change explaining the reasons. 17:36:07 Hmmm. 17:36:40 I realize that's not exactly what nirik said 17:37:11 sgallagh: not what nirik said but that was actually what I felt like suggesting :) 17:38:13 That kind of seems like threatening them... but I guess it's not a major threat 17:38:32 Does it? I can attempt to reword. 17:38:50 My feeling here is that a change like that probably should go through the Change process 17:38:57 well, I mean "if you don't change this we will make you file a change, which we could just reject" 17:39:21 nirik: As of right now, that's not the current state. 17:39:34 So it's more: "If you want to change this, please file a Change for consideration" 17:39:49 But I can try to find better language. 17:39:52 something like "If a separate /boot partition will remain a hard requirement in the Anaconda Web UI, we would like this to go through the Change Process for Fedora 40." 17:39:56 meybe? 17:39:56 ok, so perhaps rewording might be better then... 17:40:12 decathorpe: /me likes 17:40:39 decathorpe I was trying to avoid specifying a particular implementation, since they COULD opt to do this in the GTK UI too, for consistency :) 17:40:48 I guess thats ok, it still seems a bit hasty, since we haven't heard from them about it and it's not for many more months, but meh 17:41:04 sgallagh: *theoretically*, yes :) 17:41:08 I mean, there is already the web ui change right? 17:41:26 changing that in the GTK UI would also require a change proposal, no? 17:41:40 right ... but switching an UI and changing requirements for installation are orthogonal issues 17:41:50 nirik: There is, but I worry about setting a precedent that approval of that Change implicitly approves any behavioral change it causes. 17:41:58 also note that the webui is workstation only right now... 17:42:00 decathorpe said what I was trying to, better 17:42:08 webui is nobody right now 17:42:10 sure 17:42:16 well, it's still in rawhide. ;) 17:42:23 it shouldn't be 17:42:43 right. if something unanticipated slipped in that was not in the approved Change, requiring a further vote or a new CP seems reasonable 17:42:44 why not? I thought we just deferred to 40 with it? 17:42:44 anaconda team isn't even working on it since it was deferred 17:42:56 nirik: only if they were going to work on it 17:43:16 they said they weren't, so we should not have it sitting in rawhide if issues will not be addressed 17:43:18 I'm not sure what you mean there. 17:43:19 as long as it's not used by default by anything... any issue with having it in Rawhide though? 17:43:19 Proposal: A change to mandatory partition requirements in the installer should be filed as a Change for Fedora 40 17:43:29 is there some formal statement that they don't want to work on it anymore? 17:43:50 in the Workstaton WG room, we were told that as a consequence of deferral, they won't work on it anymore until F40 Beta 17:44:04 Last I heard, the plan was for them to try to arrange a way to have it as a non-default installer option in F39 GA 17:44:06 Did that go away? 17:44:13 yes 17:44:21 it got ripped out of F39 entirely 17:44:22 "until beta" - that seems like waiting too late 17:44:23 There was a lot of talk, but nothing really proposed that I know of 17:44:40 "until F40 Beta" definitely doesn't sound righ 17:44:43 *right 17:44:47 the WG was seriously unhappy when we were told that 17:45:25 I did not realize we hadn't reverted it from Rawhide after they told us that 17:45:27 perhaps we should ask them for clarification here? 17:45:34 yes 17:45:56 to be bluntly clear, I do not want people putting things in the critical path that are being ignored 17:46:08 whether it's rawhide or a branched release, that is 100% not okay 17:46:15 I know adamw has done a bunch of fixes for it in rawhide... having it there and fixing things will make f40 beta much less a fire 17:46:32 sure, we don't want that... 17:46:59 sgallagh: maybe like this: 17:47:01 Proposal: If a separate /boot partition is to become a hard requirement in the Anaconda UI, we would like this to go through the Change Process for Fedora 40 17:47:22 zbyszek: +1 17:47:31 zbyszek: +1 17:47:42 +1 17:48:01 +0.5 :) 17:48:13 * nirik breaks sgallagh's voting rules. 17:48:15 * sgallagh slaps nirik around a bit with a large trout 17:48:26 integers only? 😭 17:48:29 let's just say we round up 17:48:30 huh... apparently that command exists in the webchat 17:48:33 anyhow, I guess +1... but it seems still hasty to me. 17:48:37 nirik: physics rouding (towards even) or accountant rounding (up)? 17:48:41 sgallagh: i'm +1 on that 17:49:20 +1 17:49:24 mhayden To zbyszek's proposal or my slapping nirik with a fish? 17:49:37 sgallagh: why not both? :) 17:49:49 zbyszek: ah, didn't realize there's a term for rounding towards even 17:49:49 * mhayden doubts sgallagh uses mIRC 17:50:03 sgallagh: depends on what fish 17:50:10 mhayden Apparently that bit of ancient history was resurrected by the Libera web chat 17:50:33 👏 17:51:05 I count +6 from FESCo members 17:51:19 #agreed If a separate /boot partition is to become a hard requirement in the Anaconda UI, we would like this to go through the Change Process for Fedora 40 17:51:22 #undo 17:51:22 Removing item from minutes: AGREED by sgallagh at 17:51:19 : If a separate /boot partition is to become a hard requirement in the Anaconda UI, we would like this to go through the Change Process for Fedora 40 17:51:29 #agreed If a separate /boot partition is to become a hard requirement in the Anaconda UI, we would like this to go through the Change Process for Fedora 40 (+6, 0, -0) 17:51:44 #topic #3067 Change: Restructure Kubernetes Packages 17:51:45 .fesco 3067 17:51:46 sgallagh: Issue #3067: Change: Restructure Kubernetes Packages - fesco - Pagure.io - https://pagure.io/fesco/issue/3067 17:51:52 +1 17:52:07 I put this one on the meeting mostly because it was otherwise scheduled to be auto-rejected today, since we've ignored it for 14 days 17:52:28 oops. +1 17:52:35 unless it's really necessary, I would like to not cast my vote on this. I know *zero* things about kubernetes (other than that it's apparently a greek nautical term) 17:52:49 it's just collapsing a bunch of subpackages 17:52:58 I don't think it was worth a Change proposal 17:53:04 but that's what they did, so ehh 17:53:17 I think they wanted to get eyes on it 17:53:23 I'm also +1 17:53:36 But right now, the Change page does not actually describe the latest version of the proposal. So I'm not sure what we're actually approving. 17:53:38 i'm super +1 on this -- it was confusing to find the different packages before 17:53:43 It looks like they have a decent handle on what the Obsoletes/Provides will need to be, so I don't think anything is likely to go too awry 17:53:59 I guess "Change page + changes listed in the discussion". 17:55:34 Anyway, +1. The details can be figured out. 17:56:40 Son_Goku Can I interpret "ehh" as a +1? 17:56:50 I gave a +1 already 17:56:59 Oh, sorry. I missed that 17:57:13 ... and yet had counted it. 17:57:18 #agreed This Change is approved. (+5, 0, -0) 17:57:23 #topic Open Floor 17:59:07 Anything? 17:59:13 🦗 17:59:23 If not, I'll close out the meeting in 120 seconds 17:59:44 🪑️ 18:00:22 ⏳ 18:00:30 Oh, whoops 18:00:33 Thanks decathorpe 18:00:38 #topic Next Week's Chair 18:00:50 Who wants it? 18:01:59 Didn't somebody volunteer last week? 18:02:04 yes, i think that was me! 18:02:09 i can do next week 18:02:21 as long as you're prepared for emojis 💪 18:02:28 #action mhayden to chair the 2023-10-05 meeting 18:02:43 if IRC can handle it, then we can handle it 18:02:45 i have a standing desk, but i'll chair it nonetheless 18:03:38 Five thousand years later and we're back to hieroglyphics 18:03:48 OK, thanks for coming, folks. 18:03:56 #endmeeting