16:04:14 <jkurik> #startmeeting F28 Beta Go/No-Go meeting
16:04:14 <zodbot> Meeting started Thu Mar 29 16:04:14 2018 UTC.  The chair is jkurik. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:04:14 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:04:14 <zodbot> The meeting name has been set to 'f28_beta_go/no-go_meeting'
16:04:24 <kparal> so, now or in an hour?
16:04:25 <jkurik> #meetingname F28-Beta-Go-No-Go-meeting
16:04:25 <zodbot> The meeting name has been set to 'f28-beta-go-no-go-meeting'
16:04:37 <jkurik> #topic Roll Call
16:04:40 <kparal> I can't find any email invite. and private calendar invite says now
16:04:43 <mboddu> .hello mohanboddu
16:04:44 <zodbot> mboddu: mohanboddu 'Mohan Boddu' <mboddu@bhujji.com>
16:04:46 <jkurik> .hello jkurik
16:04:48 <zodbot> jkurik: jkurik 'Jan Kurik' <jkurik@redhat.com>
16:04:50 <lruzicka> callendar says now
16:04:52 * satellit listening
16:04:53 <jkurik> #chair nirik adamw sgallagh mboddu
16:04:53 <zodbot> Current chairs: adamw jkurik mboddu nirik sgallagh
16:04:59 <sgallagh> .hello2
16:05:00 <lruzicka> .hello2 lruzicka
16:05:00 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
16:05:03 <zodbot> lruzicka: lruzicka 'None' <lruzicka@redhat.com>
16:05:07 <lruzicka> .hello2
16:05:09 * mboddu has to take care of something, be back in few min
16:05:09 <zodbot> lruzicka: lruzicka 'None' <lruzicka@redhat.com>
16:05:36 <jkurik> I am sorry for the noise with the meeting time, however due to summer/winter time change in Europe during the last weekend, some meetings were automagicaly rescheduled
16:05:56 <kparal> .hello2
16:05:57 <zodbot> kparal: kparal 'Kamil Páral' <kparal@redhat.com>
16:06:04 <kparal> not sure if we have adamw
16:06:04 <frantisekz> one of my calendar says at 7pm, the other at 6pm, I honestly don't care at what time it'll happen at the end
16:06:06 <frantisekz> .hello2
16:06:07 <zodbot> frantisekz: frantisekz 'František Zatloukal' <fzatlouk@redhat.com>
16:06:30 <jkurik> #topic Purpose of this meeting
16:06:36 <nirik> morning
16:06:37 <jkurik> #info Purpose of this meeting is to check whether or not F28 Beta is ready for shipment, according to the release criteria.
16:06:43 <jkurik> #info This is determined in a few ways:
16:06:48 <jkurik> #info * No remaining blocker bugs
16:06:52 <jkurik> #info * Release candidate compose is available
16:06:57 <jkurik> #info * Test matrices for Beta are fully completed
16:07:01 <adamw> .hello adamwill
16:07:02 <zodbot> adamw: adamwill 'Adam Williamson' <awilliam@redhat.com>
16:07:04 <jkurik> #topic Current status
16:07:15 <jkurik> F28 RC1 & RC3 compose are ready at
16:07:22 <jkurik> http://dl.fedoraproject.org/pub/alt/stage/28_Beta-1.1/
16:07:26 <jkurik> http://dl.fedoraproject.org/pub/alt/stage/28_Beta-1.3/
16:07:35 <kparal> adamw: I was worried I'd need to chair the blocker review part. phew!
16:07:39 <jkurik> Test matrices:
16:07:45 <jkurik> https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.1_Summary
16:07:51 <jkurik> https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.3_Summary
16:08:06 <jkurik> Anyone wants to add something ?
16:08:35 <jkurik> My understanding is that we should decide whether RC1 or RC3 is more "suitable" for delivery
16:09:29 <jkurik> Our decision should be based on blockers for each of these RCs as well as on the TestMatrice coverage
16:09:34 <adamw> well, i guess first we decide if either of them are
16:09:37 <adamw> if only one is, we ship that
16:09:40 <adamw> if both are, we make a choice
16:09:45 <adamw> if neither are, we ship nothing. :P
16:09:56 <jkurik> adamw: right, thanks
16:10:16 <jkurik> #info F28 Beta RC1 & RC3 compose and test matrices are ready
16:10:22 <jkurik> #link http://dl.fedoraproject.org/pub/alt/stage/28_Beta-1.1/ - The RC1 compose
16:10:26 <jkurik> #link http://dl.fedoraproject.org/pub/alt/stage/28_Beta-1.3/ - The RC3 compose
16:10:34 <jkurik> #link https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.1_Summary - RC1 Test matrices
16:10:42 <jkurik> #link https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.3_Summary - RC3 Test matrices
16:10:50 <jkurik> Let's do at least Mini-blocker review
16:10:56 <jkurik> adamw: may I ask you please to chair the mini-blocker review ?
16:10:58 <adamw> sure
16:11:02 <jkurik> #topic Mini-Blocker Review
16:11:07 <jkurik> #link https://qa.fedoraproject.org/blockerbugs/milestone/28/beta/buglist
16:12:23 <adamw> so, we'll just do proposed and accepted Beta blockers for now.
16:12:29 <adamw> #topic (1561209) module does not properly track stream when another package Obsoletes: it
16:12:29 <adamw> #link https://bugzilla.redhat.com/show_bug.cgi?id=1561209
16:12:29 <adamw> #info Proposed Blocker, dnf, ON_QA
16:12:44 <adamw> as this is something of a corner case and the specific instance of it we know about has been worked around, -1 blocker for me
16:12:56 <sgallagh> Yes, -1 blocker from me as well
16:13:01 <frantisekz> -1 blocker
16:13:03 <adamw> note the 'criterion' here is the list of requirements at https://bugzilla.redhat.com/show_bug.cgi?id=1537253#c6
16:13:15 * nirik is still -1 as in bug
16:14:33 <jkurik> I am probably -1 on beta blocker but +1 on F28 Final
16:16:46 <lruzicka> I cant find info in the fesco meeting on this bug ... what have they been proposing?
16:16:47 <adamw> we don't need to vote for Final here
16:16:55 <adamw> lruzicka: that's what I linked to
16:17:13 <lruzicka> adamw: I know, but there is a lot of text and all I have seen was for 29 :)
16:20:33 <kparal> -1 beta blocker
16:20:52 <adamw> ok
16:21:02 <adamw> lruzicka: i linked to a specific comment, comment #6
16:21:04 <adamw> it has a list
16:21:54 <adamw> proposed #agreed 1561209 - RejectedBlocker (Beta) - as this is quite a specific corner case and the known instance has been worked around, we agree this doesn't constitute enough of an infringement of FESCo's requirements to block the Beta
16:22:09 <frantisekz> ack
16:22:19 <sgallagh> ack
16:22:19 <nirik> ack
16:22:27 <lruzicka> Rejected blocker ...
16:22:28 <lruzicka> ack
16:22:33 <jkurik> ack
16:22:46 <kparal> ack
16:23:01 <adamw> #agreed 1561209 - RejectedBlocker (Beta) - as this is quite a specific corner case and the known instance has been worked around, we agree this doesn't constitute enough of an infringement of FESCo's requirements to block the Beta
16:23:06 <adamw> #topic (1561763) KDE live image for Fedora 28 is oversize
16:23:06 <adamw> #link https://bugzilla.redhat.com/show_bug.cgi?id=1561763
16:23:06 <adamw> #info Proposed Blocker, LiveCD - KDE, NEW
16:23:48 <sgallagh> So, I'm personally of the opinion that blocking over media size for anything other than a strict physical device limitation is silly
16:24:00 <mboddu> -1 Blocker for Beta, but something to consider for Final and probably update the criteria
16:24:11 <mboddu> I agree with sgallagh
16:24:17 <adamw> we'd really need a formal agreement to amend the criteria in order to proceed here
16:24:21 * satellit claims added lots of font files
16:24:29 <adamw> (or a statement from KDE SIG that the size target had changed)
16:24:50 <adamw> if everyone agrees we should amend the criterion in at least vaguely the same direction, we can go with that
16:24:53 <jkurik> I agree with mboddu, sgallagh and nirik's comment #1
16:24:55 <frantisekz> yeah, definitely -1 blocker
16:25:16 <kparal> the easiest way forward is to move the criterion to Final
16:25:24 <kparal> I believe we can agree on that
16:25:26 <nirik> yeah, I don't think it makes sense to block things that go to usb... but of course it's still nice if they are smaller (download time, etc)
16:25:36 <jkurik> kparal: +1
16:25:53 <nirik> sure
16:26:00 <lruzicka> If we explicitly say that it is a USB image only, then no big deal
16:26:17 <sgallagh> Proposal: Move the criterion to Final and ask KDE SIG to determine the correct value
16:26:18 <adamw> kparal: i don't actually entirely, no.
16:26:21 <lruzicka> but if Live should be burnable to a CD, it should fit.
16:26:32 <nirik> you can actually get 2GB usb sticks... in bulk. But DVD media is still tons cheaper
16:26:32 <lruzicka> at least for final.
16:26:34 <adamw> kparal: i still think we should block on things being of a reasonable size for optical physical media at beta.
16:26:59 * satellit dual layer
16:27:00 <kparal> lruzicka: you can see which images are release blocking on optical media in this page: https://fedoraproject.org/wiki/Releases/28/ReleaseBlocking
16:27:03 <adamw> but, we probably shouldn't argue out the details at this point
16:27:12 <adamw> let's see if i can word a proposal that makes sense...
16:27:16 <kparal> adamw: we don't block for optical media for Beta
16:28:07 <kparal> our installation matrix says "Optical boot testing is optional at this stage." [meaning Beta]
16:28:29 <adamw> yeah, i know.
16:28:38 <adamw> i still think just going over size at beta kinda sucks.
16:28:39 <adamw> anyhow...
16:29:11 <kparal> when we don't block on it at this point, I don't understand the objection really. you would make it even stricter than it is now
16:29:52 <adamw> proposed #agreed 1561763 - RejectedBlocker (Beta) - there is general agreement that blocking on size limits not related to optical media or ISO spec size limitations at Beta is not necessary. We will work out a full agreement on how we want to revise the criterion for Beta and Final after the meeting, but there is solid consensus it should not cover this case
16:30:00 <mboddu> ack
16:30:01 <frantisekz> ack
16:30:07 <satellit> ack
16:30:09 <kparal> ack
16:30:09 <adamw> kparal: not in practice, because the *largest* size limit we have now is the size of a DVD, I believe.
16:30:11 <jkurik> ack
16:30:20 <lruzicka> ack
16:30:33 <kparal> ok, let's fine tune on test list
16:31:26 <adamw> #agreed 1561763 - RejectedBlocker (Beta) - there is general agreement that blocking on size limits not related to optical media or ISO spec size limitations at Beta is not necessary. We will work out a full agreement on how we want to revise the criterion for Beta and Final after the meeting, but there is solid consensus it should not cover this case
16:31:59 <adamw> #info of the accepted blockers, 1559680 and 1558641 are VERIFIED, meaning the fix has been confirmed in the RCs and we only need to push the packages stable
16:32:06 <adamw> #topic (1537253) Add-On Modularity
16:32:06 <adamw> #link https://bugzilla.redhat.com/show_bug.cgi?id=1537253
16:32:06 <adamw> #info Accepted Blocker, Changes Tracking, ON_QA
16:32:12 <adamw> so this is the only other outstanding accepted blocker
16:32:25 <adamw> i think we just need to affirm that we believe the Beta meets the requirements from FESCo in comments 6 and 7
16:33:41 <nirik> right
16:33:45 <sgallagh> To the best of my knowledge, with the exception of two edge cases we've discovered, this is working
16:34:07 <sgallagh> (One being the blocker we discovered earlier, the other being a UX bug I found this morning)
16:34:11 <mboddu> adamw: What about the point 6 in https://bugzilla.redhat.com/show_bug.cgi?id=1537253#c6
16:34:12 <mboddu> ?
16:34:22 <sgallagh> I didn't think it remotely severe enough to raise as a blocker
16:34:52 <sgallagh> .bug 1561816
16:34:54 <zodbot> sgallagh: Bug 1561816 – Installing two modules at the same time doesn't work - https://bugzilla.redhat.com/1561816
16:34:55 <sgallagh> (For those curious)
16:35:31 <adamw> probably worth a commonbugs note.
16:35:34 <sgallagh> mboddu: What is the question?
16:35:37 <sgallagh> adamw: Yeah
16:36:57 <mboddu> sgallagh: I remember there an issue with installing a module and when we update it, it picks from normal repo rather than module repo and I think langdon raised that issue
16:37:13 <sgallagh> mboddu: That was the BZ we discussed during the blocker review
16:37:30 <mboddu> sgallagh: Oh sorry, I missed that convo
16:37:51 * mboddu came in late few min
16:37:57 <sgallagh> It was limited to incorrect Obsoletes: handling and we worked around the only known instance for Beta
16:38:12 <adamw> ok, so
16:38:32 <mboddu> sgallagh: Okay, cool, thanks for the info
16:38:34 <adamw> proposed #info we affirm that all the requirements specified by FESCo for Beta seem to be met according to current testing
16:38:38 <adamw> patch
16:38:52 <adamw> proposed #info 1537253 - we affirm that all the requirements specified by FESCo for Beta seem to be met according to current testing
16:38:58 <nirik> ack
16:39:03 <jkurik> ack
16:39:15 <mboddu> ack
16:39:30 <sgallagh> ack
16:39:37 <frantisekz> ack
16:39:50 <lruzicka> ack
16:39:54 <adamw> #info 1537253 - we affirm that all the requirements specified by FESCo for Beta seem to be met according to current testing
16:40:00 <adamw> ok, with that, i think we're through blocker review
16:40:14 <jkurik> thanks adamw
16:40:18 <adamw> #info this completes blocker review, and there are no known outstanding blockers in RC1 or RC3
16:40:32 <jkurik> #topic Test Matrices coverage
16:40:37 <jkurik> #link https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.1_Summary - RC1 Test matrices
16:40:44 <jkurik> #link https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.3_Summary - RC3 Test matrices
16:42:13 <jkurik> looks like we have quite good coverage even for RC3
16:42:31 <adamw> yeah.
16:42:36 <adamw> so
16:43:02 <adamw> there's absolutely no reason the AD joining tests would work in RC1 but fail in RC3
16:43:18 <adamw> other than that i'm not actually seeing any missing beta coverage in rc3
16:43:21 * nirik tosses the bots a botsnack
16:43:30 <adamw> oh, i guess checksums
16:43:52 <mboddu> adamw: I did that, but I am not able to update the wiki
16:44:00 <adamw> oh, great
16:44:07 <adamw> mboddu: why not?
16:44:08 <sgallagh> adamw: Yeah, sorry I didn't have a chance to retest them
16:44:25 <sgallagh> (RE: ad join)
16:44:53 <mboddu> adamw: Not sure, when I tried to edit it:
16:44:54 <mboddu> == [[Test Results:Fedora 28 Beta 1.3 Installation|Installation]] ==
16:44:54 <mboddu> {{Test Results:Fedora 28 Beta 1.3 Installation}}
16:44:57 <mboddu> Thats all I can see
16:45:39 <adamw> mboddu: oh, you have to edit the Installation page, not the summary page.
16:46:09 <adamw> if you edit a *section* within the summary page it'll actually magically work, but if you edit the whole summary page, yeah, you're editing the summary page itself. :)
16:46:27 <mboddu> adamw: Ahhh sorry, but I found https://bugzilla.redhat.com/show_bug.cgi?id=1562146
16:46:37 <adamw> so either go to the Installation page and edit that, or click the 'edit' link next to the "Image Sanity" text.
16:46:37 * satellit https://fedoraproject.org/wiki/Test_Results:Fedora_28_Beta_1.3_Installation
16:46:45 <mboddu> But its not affecting release blocking artifacts
16:47:28 <adamw> mboddu: oh, yeah, i think i actually noticed that earlier this cycle
16:47:30 <adamw> don't recall if i filed it
16:47:33 * mboddu should change the subject of the ticket tho
16:47:41 <adamw> yeah, i did: https://bugzilla.redhat.com/show_bug.cgi?id=1497458
16:47:50 <adamw> oh no, that's an old one from dennis
16:47:56 <adamw> darn browser history misleading me. :P
16:48:14 <adamw> oh, the one i was thinking of is https://pagure.io/releng/issue/7359 .
16:48:17 <adamw> anyhow, not blocking.
16:48:32 <lruzicka> ok
16:48:36 <adamw> sooo, anyway - anyone see any missing beta coverage?
16:48:44 <mboddu> adamw: I found that but, it seems different from what I found, but we can go ahead for now
16:49:19 <jkurik> so IMO we can consider both Matrices (RC1 & RC3) as covered and move on to the decision point
16:49:20 <adamw> be good for someone to test FMW on Windows, i guess...
16:49:39 <jkurik> or anyone see some other issue with the Matrices ?
16:50:04 <kparal> nope
16:50:18 * lruzicka doesnt
16:50:18 <jkurik> #topic Go/No-Go decision
16:51:06 <jkurik> adamw: perhaps would you like to summarize the difference between RC1 and RC3 ?
16:51:26 <jkurik> just briefly....
16:51:26 <adamw> it's pretty simple
16:51:27 <nirik> systemd
16:52:07 <adamw> RC3 has systemd-238-7, RC1 has systemd-238-5.
16:52:28 <adamw> and RC3 has https://pagure.io/workstation-ostree-config/c/51df1357d419176f449a1926caa231d5cb821a9b?branch=f28 , while RC1 does not.
16:52:40 <mboddu> Dont want to confuse but RC2 has systemd-238-6 :)
16:52:46 <jkurik> the systemd-238-7 is fixing an Atomic issue, right ?
16:53:08 <adamw> they both fix Atomic issues.
16:53:17 <adamw> RC2 is right out of the running, forget RC2. :P
16:53:45 <jkurik> ok :) so what is the criterion what RC we should release ?
16:53:54 <adamw> the systemd change fixes package layering for both Atomic Host and Atomic Workstation (AIUI) - that's https://bugzilla.redhat.com/show_bug.cgi?id=1559141
16:53:55 <nirik> I think since we have good coverage, we should ship rc3...
16:54:23 <adamw> the workstation-ostree-config change fixes boot of Fedora Atomic Workstation; in RC1 it's basically DOA.
16:54:26 <mboddu> I agree with nirik, if we think we have enough coverage for RC3, then release RC3
16:54:28 <sgallagh> mboddu: The question is basically: "How nervous are we about shipping a systemd version with only a couple days' testing"
16:54:29 <adamw> both the fixes have been confirmed, I believe.
16:54:43 <adamw> sgallagh: wellllll....the systemd version in RC1 had only a couple of days of testing too.
16:54:51 <sgallagh> This is fine.
16:54:54 <adamw> =)
16:55:10 <mboddu> sgallagh: True, but systemd versions in all RC's are two days old in testing :)
16:55:13 <adamw> there is not, unfortunately, a nice clean policy for making this decision, because i'm a rebel and broke all the rules
16:55:18 <sgallagh> I'm good with RC3 though. The coverage is good
16:55:34 <mboddu> adamw: haha :D
16:56:08 <adamw> technically we should not have done RC2 and RC3, because the things they fixed were not release blocking issues. but i thought it was important to fix Atomic. though Atomic deliverables aren't technically release blocking, Atomic *is* one of our three major editions, and the reasons the images aren't release blocking are more procedural than related to how important those images are.
16:56:24 <adamw> so it seemed like a good call to try and make Atomic actually *work* in Beta.
16:56:36 <jkurik> adamw: right, good decision IMO
16:56:55 <adamw> the tradeoff, as sgallagh says, is that it means pulling in some very late systemd changes.
16:57:14 <adamw> the workstation ostree config change is pretty damn safe, it can only affect FAW, and it *clearly* made that better.
16:57:52 <mboddu> So, what is the question here, should we do more testing of systemd?
16:58:30 <adamw> what changed in systemd is, *all* the scriptlets were ported (back) from lua to sh. they were initially ported from sh to lua a few months ago in pursuit of efficiency (IIRC), but when one of them was moved from systemd to systemd-udev in systemd-238-5, this turned out to break package layering in Atomic, apparently because rpm-ostree can't deal with lua scriptlets.
16:58:51 <adamw> so if RC3 breaks anything, it's going to be those systemd package scriptlets.
16:59:22 <adamw> so, on that note: have folks tested installing, removing, updating etc. packages on top of RC3? especially packages with systemd services? has anyone noticed any scriptlet errors or things like that?
16:59:37 <sgallagh> I think it was probably more about eliminated a dependency loop on /bin/sh because RPM includes an embedded Lua interpreter
17:00:00 <adamw> sgall:
17:00:01 <adamw> Convert file trigger scripts to lua
17:00:01 <adamw> 
17:00:01 <adamw> At least the %filetriggerpostun script can be invoked hundreds of
17:00:02 <adamw> times during an upgrade, so it makes sense to optimize it a bit.
17:00:04 <adamw> that was the commit message.
17:00:10 <adamw> note, 'optimize'
17:00:11 * sgallagh shrugs
17:00:57 <jkurik> proposed #agreed As we have good coverage for the RC3, the Go/No-Go decision is SHIP the RC3 as F28 Beta on 2018-Apr-03
17:01:29 <adamw> we only have a couple of votes so far...
17:01:45 <jkurik> ... or anyone feels like we should ship the RC1 (due to the systemd) ?
17:01:55 <kparal> ack as proposed
17:02:01 <nirik> ack as proposed
17:02:05 <adamw> i think i'd vote for RC3, though. i am very *slightly* worried some subtle issue or typo in one of the ported scripts will show up as people do more updates and things, but overall i think fixing Atomic is a substantial benefit.
17:02:08 <adamw> so, ack
17:02:28 <sgallagh> ack
17:02:32 <adamw> (i guess we should note that the reason RC2 is out of the running is, there were some subtle typos in the ported scripts that broke them. :>)
17:03:00 <lruzicka> ack
17:03:02 <jkurik> mboddu: is releng able to release the RC3 on 2018-Apr-03 even Friday is off ?
17:03:08 <mboddu> ack as proposed and if we find something(hope not), we will ship an update
17:03:18 <mboddu> jkurik: Yes
17:03:25 <jkurik> mboddu: thanks
17:03:34 <jkurik> #agreed As we have good coverage for the RC3, the Go/No-Go decision is SHIP the RC3 as F28 Beta on 2018-Apr-03
17:03:48 <jkurik> #action jkurik to publish the Go/No-Go result
17:03:55 <jkurik> #topic Open floor
17:03:59 <jkurik> anything else to discuss today on this meeting ?
17:04:09 * nirik has nothing.
17:04:17 * mboddu has nothing
17:04:22 <sgallagh> \o/
17:05:15 <jkurik> #info For the record: RC2 was not considered for the release as  there were some subtle typos in the ported scripts
17:05:19 <adamw> yay
17:05:22 <adamw> thanks everyone
17:05:27 <kparal> \o/
17:05:27 <clime> hello
17:05:31 <jkurik> thanks everyone
17:05:37 <jkurik> #endmeeting