15:00:06 <nirik> #startmeeting RELENG (2023-05-02)
15:00:06 <zodbot> Meeting started Tue May  2 15:00:06 2023 UTC.
15:00:06 <zodbot> This meeting is logged and archived in a public location.
15:00:06 <zodbot> The chair is nirik. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
15:00:06 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:06 <zodbot> The meeting name has been set to 'releng_(2023-05-02)'
15:00:08 <nirik> #meetingname releng
15:00:08 <zodbot> The meeting name has been set to 'releng'
15:00:14 <nirik> #chair nirik sharkcz pbrobinson phsmoura dustymabe jednorozec
15:00:14 <zodbot> Current chairs: dustymabe jednorozec nirik pbrobinson phsmoura sharkcz
15:00:18 <nirik> #topic init process
15:00:30 <nirik> jednorozec said he would be 10-15m late. ;)
15:01:02 <nirik> I didn't have too much today... We are going to have a ticket triage/closing meeting after this one (at 16UTC).
15:01:32 <nirik> And I would like to move koji builders/hubs to f38 soon. I'm happy to do that and document it more, or work with others to do it.
15:01:37 <patrikp[m]> Hello everyone. 👋
15:02:06 <nirik> Hey patrikp!
15:03:24 <patrikp[m]> May I ask where the ticket cleanup meeting will take place? This room?
15:03:46 <mkonecny> Hi everyone
15:04:27 <nirik> Yeah, I think this room. ;)
15:04:42 <nirik> hey Zlopez
15:06:13 <mkonecny> I wanted to just look what is being discussed here
15:06:31 <nirik> well, not much yet... waiting for jednorozec :)
15:10:05 * nirik wonders if he can go feed cats real quick. ;)
15:10:15 <nirik> back in a min or 10 or 5.
15:12:10 <mkonecny> I will go for a tea
15:19:27 <nirik> ok, back.
15:19:42 <mkonecny> Me too
15:22:26 * jednorozec[m] is back
15:22:35 <jednorozec[m]> so
15:22:40 <jednorozec[m]> what do we have for today
15:23:59 <jednorozec[m]> not much on my plate for today, but
15:24:09 <jednorozec[m]> there is a new change request with another new artifact
15:24:12 <jednorozec[m]> .releng 11411
15:24:12 <zodbot> jednorozec[m]: Issue #11411: F39 Change: Fedora Onyx - releng - Pagure.io - https://pagure.io/releng/issue/11411
15:24:36 <nirik> yeah... will require updating a bunch of things... if/when approved.
15:24:52 <jednorozec[m]> yeah
15:25:06 <jednorozec[m]> I started draft on adding new artifacts for the docs
15:25:30 <nirik> yeah, I think thats a good plan... although things change quickly sometimes... so we will have to try and keep it up to date.
15:26:48 <jednorozec[m]> well i hope we will keep the new docs up to date with each new release
15:28:15 <nirik> oh, I had 2 more things i just remembered
15:28:33 <jednorozec[m]> go for it
15:29:27 <nirik> first, theres a private releng ticket ( 11164 ). I have a draft of an announcement there... I'd like to send that out today if it looks ok, or get feedback on it if it's not clear.
15:29:49 <nirik> I can discuss that out of band if needed. ;)
15:29:54 <nirik> and second...
15:30:10 <nirik> I made a PR for a announcement about inactive packager removals.
15:30:11 <jednorozec[m]> oh that one
15:30:26 <nirik> https://pagure.io/releng/pull-request/11408
15:30:42 <nirik> I'd like to finalize that oen and send it too. (or you can... ;)
15:31:11 <nirik> it needs the output from the infra ticket where it says what packages were orphaned, etc.
15:32:24 <jednorozec[m]> +1 on both
15:32:42 <jednorozec[m]> the private ticket was little bit scary
15:32:58 <nirik> Cool. Shall I send both? or do you want to send one or the other?
15:33:02 <nirik> yes. ;(
15:33:48 <jednorozec[m]> well you rolled out the fix, i think you should take the fame
15:34:27 <nirik> ha. ok, I will send them out... see what flames I get.
15:35:09 <jednorozec[m]> ok
15:35:55 <jednorozec[m]> did you see comments by mike on the policy thing? https://pagure.io/fedora-infra/ansible/pull-request/1365
15:36:08 <nirik> jednorozec: on f38 upgrades. Did you want me to just do it and document, or did you want to try and do it together or ? I know timezones aren't good as always.
15:37:08 <jednorozec[m]> so it should be just change the repo and dnf distrosync?
15:37:10 <nirik> yes, I did. I guess that PR might be ok to push out? I can look again... we can try before upgrading? or after?
15:37:58 <jednorozec[m]> nirik: I dont mind doing it alone, just wasnt sure about when
15:38:28 <nirik> well, it's more complicated than that to do things without outages... and virthosts need updating/rebooting. At least the 'simple' path (just upgrading in place). There's also the 'hard' path which is reinstalling everything from the ground up. That takes a lot longer/more effort. ;)
15:38:47 <mkonecny> I looked at that PR today and noticed that nirik said it waits for fixes in koji
15:38:51 <mkonecny> Are there in place?
15:38:56 <nirik> we should also do staging and make sure things are looking ok on 38 first...
15:39:13 <jednorozec[m]> so
15:39:27 <jednorozec[m]> I will update some in stg in my morning tmrw
15:39:44 <nirik> jednorozec: staging should just be 'dnf --releasever 38 distro-sync -y' on builders and koji01.stg... and rebooting. You want to just do that in the next few days and confirm it's ok?
15:39:50 <nirik> yeah. ;)
15:40:10 <nirik> Zlopez: yes, the fix is the private ticket I mentioned... and I deployed friday.
15:40:15 <jednorozec[m]> ack, I will test it on stg and if things work prod is next
15:40:28 <jednorozec[m]> tmrw will be the upgrade stg day :)
15:40:42 <nirik> +1
15:40:42 <mkonecny> nirik: Thanks for info :-)
15:41:28 <nirik> I'm not sure if/when would be good to roll out that pr... it will be distruptive. Perhaps when we do f38 updates? we can land it right before?
15:41:40 <nirik> and apply on updated machines as we update them?
15:41:58 <nirik> but it has a hub part too... so, hum
15:42:06 <nirik> I think perhaps we should wait and do an outage.
15:42:26 <nirik> next week or week after we probibly need to do a mass update/reboot cycle anyhow.
15:43:23 <jednorozec[m]> lets update first
15:43:37 <jednorozec[m]> once we have things on 38 lets do disruptive changes
15:43:48 <nirik> yeah, sounds reasonable.
15:44:35 <nirik> Lets see how nicely staging goes and plan from there. ;)
15:46:21 <nirik> 15min to ticket review... I think here?
15:46:31 <jednorozec[m]> yes here
15:46:57 <jednorozec[m]> Lets end this sooner so we have time to grab water/coffee/food
15:46:57 <mkonecny> I have one thing as well, the shared pungi config for Bodhi
15:47:08 <nirik> ah yeah. ;(
15:47:13 <jednorozec[m]> oh
15:47:27 <mkonecny> I saw that nirik reverted the commits as they caused issues again :-/
15:47:37 <mkonecny> Thanks for taking care of that
15:47:38 <jednorozec[m]> is it working? I am getting some weird messages from stg bodhi
15:47:55 <nirik> yeah. We need the new files to be templates too, not just files... since they need to expand {{ release }} type things.
15:48:12 <nirik> it's not working. I reverted it. ;)
15:48:26 <mkonecny> I will look at it this week and adjust that
15:48:37 <mkonecny> Hopefully on third time it will work
15:48:41 <nirik> We got things like: pungi.compose.status.change -- pungi-koji compose of [[ release.id_prefix.title() ]]-36-updates-20230429.0 started https://kojipkgs.fedoraproject.org/compose/updates/[[ release.id_prefix.title() ]]-36-updates-20230429.0
15:49:01 <nirik> we will get there someday! :)
15:49:27 <jednorozec[m]> baby steps
15:49:41 <nirik> should we just keep this meeting open? or close and re-open a ticket review one?
15:50:07 <jednorozec[m]> keep it open I think
15:50:16 <jednorozec[m]> we will have logs from the ticket triage in releng weekly logs
15:50:29 <nirik> sure.
15:50:42 * nirik goes to get some coffee.
15:50:50 * jednorozec[m] goes to hunt some food
15:55:20 * nirik is back
15:55:29 <nirik> hopefully we have mboddu joining us. ;)
15:56:46 <mkonecny> It would be nice to see him again :-)
15:58:59 <nirik> yes!
16:00:21 <jednorozec[m]> so they have cats here where I sleep in Brno
16:00:28 <jednorozec[m]> and they are fighting me for food
16:00:32 <jednorozec[m]> anyhow
16:00:42 <jednorozec[m]> here is the list of tickets https://pagure.io/releng/issues?status=Open&order_key=last_updated&order=asc
16:00:56 <mboddu_> Yes, I am here
16:01:15 <mkonecny> Hi mboddu_
16:01:18 <mboddu_> Are we meeting here or on gmeet?
16:01:25 <nirik> ey mboddu_!
16:01:31 <jednorozec[m]> mboddu_: here for the logs
16:01:36 <nirik> here we thought... so we can have record...
16:01:51 <mboddu> Ack
16:02:50 <mboddu> Take it away jednorozec[m]
16:02:58 <nirik> so, lets start at oldest and do them one at a time? or should we focus on ones we can get info from mboddu_ while he's here?
16:03:52 <mboddu> I dont think there is an easy way to identify tickets that require my input, lets start from the oldest
16:04:08 <mboddu> We can schedule another meeting if needed
16:04:29 <nirik> ok.
16:04:33 <jednorozec[m]> .releng 8646
16:04:35 <zodbot> jednorozec[m]: Issue #8646: Create Fedora LiveOS images using plain squashfs root filesystem - releng - Pagure.io - https://pagure.io/releng/issue/8646
16:04:44 <nirik> so, I was looking at this...
16:04:58 <nirik> the koji thing we said it was blocked on is still pending/not done.
16:05:21 <nirik> Proposal: update ticket asking reporter what the status is and if it's still desired?
16:05:34 <nirik> (since they would be better pinging on upstream stuff than us)
16:05:43 <jednorozec[m]> do you want me to do it or will you?
16:06:29 <nirik> if you could that would be great, but I can if you prefer?
16:06:37 <mboddu> (My 2 cents is, its good to use overlayfs, but it depends on koji to prioritize that work)
16:06:42 * jednorozec[m] writing comment
16:06:49 <jednorozec[m]> nirik: pick another ticket
16:07:07 <nirik> .releng 9392
16:07:08 <zodbot> nirik: Issue #9392: explore / investigate new-updates-sync for rawhide/branched composes - releng - Pagure.io - https://pagure.io/releng/issue/9392
16:08:14 <nirik> so... not sure on this one. I don't think anything was changed...
16:09:27 <nirik> proposal: I guess ping on it and see if it's still desired?
16:10:02 <mboddu> Yeah, not sure if we have to do anything
16:10:12 <nirik> I can update the ticket and we can pick the next
16:10:56 <nirik> updated
16:11:00 <nirik> .releng 7498
16:11:01 <zodbot> nirik: Issue #7498: Support on-the-fly tarball generation in Koji - releng - Pagure.io - https://pagure.io/releng/issue/7498
16:11:42 <nirik> so this is a long drawn out discussion...
16:12:08 * jednorozec[m] looks at the coments
16:12:35 <nirik> it's... long and many people have anges.
16:12:36 <nirik> angles
16:13:15 <nirik> IMHO, this is a good example of a discussion thats bad to have in a ticket. I mean we aren't even sure what we want to implement here.
16:13:36 <nirik> There was a group that formed to figure this out.
16:13:38 <mboddu> +1 to that
16:13:57 <jednorozec[m]> did they figured it out?
16:14:41 <nirik> Ha. I am trying to find the info. They formed up and had some meetings I know.
16:14:51 <mboddu> Not that I know of, there was an idea but nothing was implemented
16:15:00 <nirik> https://docs.fedoraproject.org/en-US/ci/source-git/
16:15:52 <nirik> but I thought there was a sig / more
16:16:21 <nirik> https://fedoraproject.org/wiki/SIGs/Source-git
16:16:36 <nirik> looks kinda inactive.
16:17:13 <nirik> anyhow, proposal: close ticket and say: hey, discuss this on the list / with the source git sig / fesco and once we figure out what we want we can work on implementing it?
16:17:17 <mboddu> They did some stuff in centos stream, but nothing (maybe started) in Fedora world
16:17:20 <nirik> also the kernel is already kinda doing this
16:18:03 <jednorozec[m]> +1 for closing
16:18:27 <jednorozec[m]> since you have been part of that can you close it nirik ?
16:18:38 <jednorozec[m]> *discussion
16:18:53 <nirik> yep. Working on comment now.
16:20:08 <nirik> k, done
16:20:14 <nirik> next?
16:20:25 <jednorozec[m]> .releng 6746
16:20:26 <zodbot> jednorozec[m]: Issue #6746: Produce a slimmed-down compose whenever certain packages appear in an update - releng - Pagure.io - https://pagure.io/releng/issue/6746
16:20:48 <jednorozec[m]> and
16:20:51 <jednorozec[m]> .releng 7428
16:20:53 <zodbot> jednorozec[m]: Issue #7428: Anaconda & bodhi update improvements - releng - Pagure.io - https://pagure.io/releng/issue/7428
16:20:56 <jednorozec[m]> they are related
16:21:02 <nirik> I know mboddu_ made a odcs thing around this?
16:21:05 <nirik> but...
16:21:14 <mboddu> Yes, this is what started fmc
16:21:26 <nirik> I'd propose we just close them, as openqa already did this. Unless there's some other need for them?
16:21:57 <jednorozec[m]> I dont think we need them in other places
16:22:18 <mboddu> There is still a need for it, but that need can be resolved using fedora ci, bodhi can call an odcs compose
16:22:44 <mboddu> There should be some documentation around it
16:22:46 * mboddu checking
16:23:49 <nirik> well, openqa already builds the iso... so... perhaps anything else that needed it could just use that?
16:23:56 <mboddu> I am not able to find it, but jkaluza should have it
16:24:57 <mboddu> ^ I did a quick look due to time
16:25:16 <jednorozec[m]> jkaluza is no longer maintaining odcs
16:25:43 <nirik> ok, so then, perhaps we close 6746 and update 7428 asking if there's any need for more than openqa already does?
16:25:52 <mboddu> +1
16:26:06 * jednorozec[m] writes comments
16:26:41 <nirik> cool.
16:27:04 <nirik> .releng 7388
16:27:11 <nirik> is also kinda of related.
16:27:15 <zodbot> nirik: Issue #7388: predicting dependency failures in image builds - releng - Pagure.io - https://pagure.io/releng/issue/7388
16:28:01 <nirik> this can be close IMHO. We have openqa/ci thats running, and if anything in the critpath has broken deps... it will catch it.
16:28:39 <mboddu> nirik: Just wondering if we can add some sort of tmt testing to composes? Will it be useful?
16:29:40 <nirik> not sure. openqa already does a ton of testing on em...
16:30:43 <mboddu> Okay
16:32:10 <nirik> next? or wait a min to catch our breath. ;)
16:32:11 <jednorozec[m]> +1 for closing
16:33:02 <nirik> I close 7388.
16:33:38 <jednorozec[m]> .releng 7994
16:33:39 <zodbot> jednorozec[m]: Issue #7994: Inform module maintainers when their module/stream is going to EOL - releng - Pagure.io - https://pagure.io/releng/issue/7994
16:34:01 <jednorozec[m]> I remeber working on this
16:34:02 <mboddu> Simply put, +1 to close it
16:34:12 <nirik> ha, you had a script... 3 years ago. ;)
16:34:35 <jednorozec[m]> but its tricky because I had some discussions with module maintainers, and most of the time they do not know EOL
16:34:55 <jednorozec[m]> fetching the EOl form PDC is nothing
16:34:56 <nirik> yeah, I am inclined to agree... I mean, it might be nice, but also, another thing querying pdc is... not good.
16:35:08 <nirik> and we have lived without it till now?
16:35:15 <jednorozec[m]> exactly
16:35:16 <jednorozec[m]> :)
16:35:21 <mboddu> ^ exactly my reason to close it
16:35:26 <nirik> but if you want to finish it...
16:35:42 <jednorozec[m]> actually people requsted atleast 3 times to set EOL on module to EOL of rawhide ...
16:36:09 <jednorozec[m]> for those who do not know that is 2222 :)
16:36:16 <jednorozec[m]> so lt me close this
16:36:21 <mkonecny> I would just ask if this is still something they want to have before closing it
16:36:25 <nirik> +1
16:36:47 <jednorozec[m]> Zlopez: do I hear your keyboard writing a comment there?
16:36:53 <nirik> well, do we know when they are? I don't know that anyone who might use it is on the ticket
16:36:55 * jednorozec[m] hides
16:37:07 <nirik> s/when/who/
16:37:20 <mkonecny> jednorozec: I can write it
16:37:23 <nirik> I mean, mohan filed it...
16:37:29 <jednorozec[m]> heh
16:37:33 <mboddu> ^^ also another ticket that will be lost until we come back to it
16:37:33 <jednorozec[m]> just realized that
16:37:36 <jednorozec[m]> :D
16:37:38 <nirik> and I see no one I associate with modules on it
16:37:47 * jednorozec[m] closes
16:37:50 <mkonecny> OK, then we can close it :-D
16:37:57 <nirik> I'd say just close.
16:38:08 <mboddu> I didn't even open the link, I saw the title and was like "close it". Didn't realize I opened it :D
16:38:31 <nirik> ha
16:39:16 <jednorozec[m]> .releng 8496
16:39:17 <zodbot> jednorozec[m]: Issue #8496: Some file provides are missing from primary.xml - releng - Pagure.io - https://pagure.io/releng/issue/8496
16:39:42 <nirik> IMHO, close and say 'if there's still something to do let us know or file upstream in createrepo_c' ?
16:40:00 <mboddu> +1
16:40:02 <nirik> again... it's not been an issue for 2 years.
16:40:40 <mkonecny> +1
16:40:41 * jednorozec[m] closes
16:42:13 <jednorozec[m]> .releng 8929
16:42:14 <zodbot> jednorozec[m]: Issue #8929: When orphaning packages, keep the original owner as co-maintainer - releng - Pagure.io - https://pagure.io/releng/issue/8929
16:42:23 * nirik reads
16:42:45 <jednorozec[m]> we should link this to pagure upstream
16:43:06 <nirik> well, hang on...
16:43:22 <nirik> I think this could just be a change in our give-package script?
16:43:43 <nirik> instead of just giving the package, we should give it and then make sure the old maintainer is still admin?
16:43:45 <mboddu> It depends, what script is being used for unresponsive maintainers?
16:44:18 <nirik> there's a seperate one for that... orphan-all-packages I think?
16:44:36 <jednorozec[m]> so
16:44:37 <nirik> but yeah, I guess it depends on the use case.
16:44:46 <jednorozec[m]> all our scripts are calling the same pagure api endpoint
16:44:53 <mboddu> Ben does run these things.... (an awkward pause)
16:45:09 <jednorozec[m]> if we change it in pagure we do 1 change for all the orphaning procedures
16:45:30 <nirik> but in some cases we don't want the old maintainer still admin.
16:45:44 <nirik> (ie, if we are removing them from packager as inactive or something)
16:45:55 <nirik> proposal:
16:45:56 <jednorozec[m]> riht
16:46:14 <jednorozec[m]> let me assign this to myself nad I will read throu it again
16:46:17 <nirik> update ticket and ask to gather a list of what processes we need to change?
16:46:42 <nirik> I think we should figure out the various paths and fix only the one(s) we need to.
16:46:51 <jednorozec[m]> yup
16:46:55 <nirik> Or perhaps even it's not needed anymore... not sure.
16:47:13 <jednorozec[m]> I still think this will be better handled in pagure, but maybe not
16:47:47 <jednorozec[m]> I will take the ticket
16:47:52 <mboddu> I think its still needed, but one another key aspect to check before assigning it to co-maintainer is to check whether that co-maintainer is active or not
16:47:55 <nirik> it might be (although... pagure isn't too active... so if we just file there, likely nothing will happen unless we do it)
16:48:53 <jednorozec[m]> if the change will be in pagure I can do it, there is a couple of changes that require pagure release anyway
16:48:53 <nirik> well, we don't want to assign to co-maintainer, just make the current main admin not the current main admin. :)
16:49:54 <nirik> but yeah, needs more sorting out of all the cases.
16:49:54 <nirik> lets do one more, then I need to get coffee before fesco. ;)
16:50:11 <mboddu> +1
16:50:17 <jednorozec[m]> .releng 8478
16:50:18 <zodbot> jednorozec[m]: Issue #8478: Retired packages should close rawhide bugzilla as WONTFIX or EOL - releng - Pagure.io - https://pagure.io/releng/issue/8478
16:50:34 <jednorozec[m]> is this already done in toddlers?
16:50:41 <mboddu> (FYI, I am happy to attend few more of these meetings if needed, but not at regular releng meeting time due to a conflict)
16:50:57 <mboddu> This needs scripting to fedpkg retire
16:51:03 <nirik> I'm not sure if it's done or not...
16:51:11 <mboddu> Or toddlers is a good idea
16:51:28 <mkonecny> Let me check the toddlers
16:51:42 <nirik> and... I argued in the ticket this wasn't too useful to do. ;)
16:52:03 <nirik> but I suppose there could be some small number of bugs it would be good to close.
16:52:25 <mkonecny> No, we don't have any toddler like that
16:53:11 <mboddu> Well, my take is sometimes its needed, sometimes its not, if we are closing them when retired, maybe reopening and assigning to new owner when unretired is also important
16:53:21 <mboddu> So that those tickets wont be lost in abyss
16:53:57 <nirik> well, if we reopen... that would be potentially tons of fixed/already dealt with bugs.
16:54:13 <nirik> but yeah. So, perhaps we just update this one to discuss more?
16:54:21 * nirik goes for cofee
16:54:26 <nirik> coffee even
16:54:30 <mboddu> I mean, the one's that got closed due to retirement
16:55:42 * jednorozec[m] is querying bugzilla
16:55:57 <mboddu> When unretired, have the script to check for "Closed -> Wont Fix" and then check the comments why it got closed, like "closed due to retirement of the pkg", reopen just those bugs
16:56:49 <mboddu> Its possible with bugzilla api, afaik
16:57:57 <nirik> thats kinda complex... I suppose it's possible tho.
16:58:05 <nirik> again, also we have lived without this so far. ;)
16:58:40 <mboddu> I am more than happy to help there, esp with toddlers, it will make these things easier
16:58:41 <jednorozec[m]> so who is writing the comment for todays last ticket?
16:58:49 <mboddu> I can take it
17:00:32 <mboddu> Well, my last comment still holds https://pagure.io/releng/issue/8478#comment-688925 the current discussion
17:00:58 <jednorozec[m]> heheh
17:01:22 <jednorozec[m]> thank you all for attention and input.
17:01:33 <jednorozec[m]> #endmeeting
17:01:42 <jednorozec[m]> ha
17:01:43 <jednorozec[m]> I didnt started it
17:01:46 <mboddu> Maybe I will take a look when I get some time on this ticket, giving back to the community
17:01:47 <mboddu> :)
17:02:24 <mboddu> Thank you jednorozec[m], nirik and mkonecny for running this meeting
17:03:30 <nirik> thanks everyone!
17:03:32 <nirik> #endmeeting