16:00:16 <jednorozec> #startmeeting RELENG (2023-06-13)
16:00:16 <zodbot> Meeting started Tue Jun 13 16:00:16 2023 UTC.
16:00:16 <zodbot> This meeting is logged and archived in a public location.
16:00:16 <zodbot> The chair is jednorozec. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:00:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:16 <zodbot> The meeting name has been set to 'releng_(2023-06-13)'
16:00:16 <jednorozec> #meetingname releng
16:00:16 <jednorozec> #chair nirik sharkcz pbrobinson phsmoura dustymabe jednorozec
16:00:16 <zodbot> The meeting name has been set to 'releng'
16:00:16 <zodbot> Current chairs: dustymabe jednorozec nirik pbrobinson phsmoura sharkcz
16:00:16 <jednorozec> #topic init process
16:00:41 <nirik> good morning everyone.
16:03:36 <jednorozec> hello nirik, how are things over there?
16:04:13 <jnsamyak> Hello
16:05:19 <nirik> not too bad...
16:06:17 <jednorozec> let get started
16:06:37 <jednorozec> so there was one new sidetag request today for perl
16:06:49 <jednorozec> .releng 11474
16:06:50 <zodbot> jednorozec: Issue #11474: F39 System-wide change: Perl 5.38 - releng - Pagure.io - https://pagure.io/releng/issue/11474
16:07:06 <jednorozec> I have created signing pr for the tag https://pagure.io/fedora-infra/ansible/pull-request/1482
16:08:03 <nirik> looks ok from a quick glance. I can push it if you like.
16:08:10 <jednorozec[m]> huh the bridge is extreamly slow swithing to matrix
16:08:39 <nirik> it's been having issues these days.
16:09:02 <jednorozec[m]> yeah
16:09:50 <nirik> ah, I see one issue with the PR, will add comment.
16:11:30 <nirik> I didn't have much today, just update on flatpak building, dnf5/mock changes and then perhaps ask about status on various in progress things. ;)
16:12:47 <jednorozec[m]> I do not have much myself, feel free to shoot questions
16:14:19 <nirik> So, on flatpaks... I finally merged that PR to move koji scm permissions from every builder to the hub. Unfortunately, that broke flatpak building because the koji-containerbuild we had was too old. ;( otaylor was able to backport the needed patches tho, and I deployed that one and it's working again.
16:15:19 <nirik> On dnf5/mock, I will push those mock changes today I guess... basically we need to override dnf for mock so it keeps using the dnf4 version for builds. dnf5 wants to land by default in rawhide, but it's not ready to do build stuff yet, so we want to keep building with older dnf for now.
16:15:41 <nirik> (thats https://pagure.io/releng/issue/11446 )
16:16:28 <jednorozec[m]> makes sense
16:17:03 <jednorozec[m]> in whitch role we store the moc configs?
16:17:07 <jednorozec[m]> *mock
16:17:52 <nirik> well, the site defaults are under roles/koji_builder/
16:18:14 <nirik> but also koji generates mock configs, so most of it's in koji... but we can override some things per site wide
16:19:29 <nirik> .ticket 11385
16:19:31 <zodbot> nirik: An error has occurred and has been logged. Check the logs for more information.
16:19:48 <nirik> oh right.
16:19:51 <nirik> .releng 11385
16:19:53 <zodbot> nirik: Issue #11385: Silverblue aarch64 installer image compose always fails on F38 and Rawhide - releng - Pagure.io - https://pagure.io/releng/issue/11385
16:20:18 <nirik> making a f38 silverblue aarch64 installer image still on your list jednorozec?
16:21:04 <jednorozec[m]> Oh I totally forget about it, but now i remember why.
16:21:28 <nirik> There was some more discussion about it yesterday... people asking if we could make it, etc.
16:21:55 <jednorozec[m]> so I wasnt sure how to do that should I just user koji imagebuild with right params and tag?
16:22:28 <jednorozec[m]> or is there something I am missing?
16:22:45 <nirik> yeah, I'd say take the orig config for the one that failed... and tweak it to not be exactly the same and then run that?
16:22:56 <jednorozec[m]> right
16:23:30 <jednorozec[m]> I think I have the comand somewhere
16:23:52 <jednorozec[m]> but Yeah I will run the image build today
16:24:32 <nirik> there should be a .cfg file under the compose area that you can just pass koji image-build...
16:24:37 <nirik> (well, after modifying)
16:25:33 <jednorozec[m]> I started modifying the params and than something else showed up
16:26:19 <jednorozec[m]> There is this one I wanted to talk about
16:26:21 <nirik> I can't seem to find it off hand, but yeah...
16:26:22 <jednorozec[m]> .releng 11460
16:26:23 <zodbot> jednorozec[m]: Issue #11460: find_unblocked_orphans.py generates false positive orphaned packages (due to 'or java-1.8.0-headless') - releng - Pagure.io - https://pagure.io/releng/issue/11460
16:27:01 <jednorozec[m]> so this lives in our repo and seems to be abandoned for some time
16:27:26 <jednorozec[m]> maybe we could move this into toddlers and make it publish the report somewhere
16:27:41 <nirik> Might be a lot of noise...
16:27:55 <jednorozec[m]> the script is really complex and I dont think it has to be
16:27:57 <nirik> but might be possible.
16:28:28 <nirik> I think first step might be talking to churchyard... I'm sure he has some good ideas on how it could be better run/used.
16:29:16 <nirik> partly tho the complaint in that ticket is just how it's reported... it's just saying that if X was removed, Y would break and his Z would be broken by that
16:29:53 <jednorozec[m]> So I looked into the sources
16:29:56 <nirik> I'm not sure how better to report it... it might need to be a web app more than a static email?
16:30:09 <jednorozec[m]> its about 900 lines of code
16:30:14 <nirik> (not that I am saying we should deploy another app :)
16:30:54 <jednorozec[m]> heh we need more apps and more layers of technology!
16:31:50 <nirik> I wonder if content resolver could be of use/a good idea here.
16:32:17 <nirik> I've not really used/looked at it much, but it sort of does this.
16:32:41 <jednorozec[m]> content resolver?
16:32:59 <nirik> https://tiny.distro.builders/views.html
16:33:10 <nirik> it's what eln and centos stream use
16:33:16 <nirik> and probibly rhel somewhere.
16:33:42 <nirik> but basically it allows you to mark packages as wanted or unwanted, and it computes deps
16:34:19 <jednorozec[m]> yeah
16:34:26 <nirik> so if it could handle fedora (no idea how it scales), orphaned packages could be marked unwanted...
16:34:29 <jednorozec[m]> tht looks like it might be usefull
16:34:34 <nirik> but just a random thought.
16:36:18 <nirik> replacing the script with something simpiler might be quicker. Anyhow, I'd suggest we talk to churchyard as he's been running that for a while.
16:38:24 <nirik> Do you want to do that? or want me to? or ?
16:38:36 <jednorozec[m]> I can do that
16:38:45 <jednorozec[m]> I bet we will meet in person at devconf
16:39:07 <nirik> ah yes, cool. ;)
16:39:28 <nirik> Last thing I had was to talk about openh264 stuff... whats status and next steps there?
16:40:09 <jednorozec[m]> huh, id we got a response from adrian last week?
16:40:23 <jednorozec[m]> the chanell names are different over here than on irc
16:40:27 <nirik> Oh nope. I'll send email.
16:40:39 <nirik> will cc you and samyak
16:40:48 <jednorozec[m]> cool
16:41:31 <jednorozec[m]> but files are in place, adding epel9 to mm oh and the epel release
16:41:39 <jednorozec[m]> I bet I had a PR for that
16:41:59 <nirik> was this repo multilib?
16:42:04 <jednorozec[m]> nope but have the new repo file in my fork
16:42:07 <jednorozec[m]> let me open it
16:42:33 <jednorozec[m]> so fedora repos are and where but there is no 686 for epel9
16:43:32 <nirik> yeah, right.
16:43:41 <nirik> so, we can close the ticket about fedora ones then?
16:43:52 <nirik> https://pagure.io/releng/issue/11425
16:43:55 <jednorozec[m]> yup
16:44:15 <jednorozec[m]> neal pinged the perspon but no reponse in the ticket
16:44:20 <nirik> cool. ;) I'll let you do the honors.
16:46:42 <nirik> I don't think I have anything else off hand.
16:46:57 <jednorozec[m]> #topic Open Floor
16:47:47 <sharkcz> hi guys, one FYI thing
16:48:04 <jednorozec[m]> hi
16:48:04 <nirik> Hey sharkcz. :)
16:48:13 <sharkcz> latest btrfs-progs is broken on s390x, but seems it's not causing issues in the composes
16:48:33 <sharkcz> https://bugzilla.redhat.com/show_bug.cgi?id=2214522, bisected and the bad commit identified
16:48:36 <nirik> ☹️
16:48:47 <sharkcz> now it's for upstream to fix it
16:49:58 <sharkcz> EOM :-)
16:50:19 <nirik> ok, cool. say, while you are here... ;) Any updates on https://pagure.io/fedora-infrastructure/issue/11122 (s390x test instance machine for maintainers) ? also... we have a old power8 box loaned from ibm still sitting in our iad2 datacenter. ;( If you have any ideas on who we can contact to ship it back that would be awesome.
16:50:43 <nirik> (I tired some folks, but emails bounced)
16:50:51 <sharkcz> sorry, no update for the new test machine
16:51:13 <nirik> ok, no worries.
16:51:16 <sharkcz> I will have some contact for shipping p8 back, will forward it to you
16:51:26 <cmurf> sharkcz: yeah we didn't build 6.3 so it's possible the regression isn't new in 6.3.1
16:51:32 <nirik> That would be great!
16:51:52 <sharkcz> cmurf: it's in 6.3.1, 6.3 was OK
16:52:17 <cmurf> sharkcz: interesting, could you mention that in the upstream bug?
16:52:23 * nirik goes to get more coffee before fesco meeting in a few min.
16:52:25 <sharkcz> yup, all commented there
16:52:40 <sharkcz> including the bad commit
16:55:03 <mhroncok> jednorozec: no defconf for me this year, sorry
16:55:34 <jednorozec> mhroncok, hopefully you have better plans ;)
16:55:45 <jednorozec> while you are around
16:56:15 <jednorozec> How often do you run the find_unblocked_orphans script? and do you do anything with the output?
16:56:16 <mhroncok> *devconf
16:56:56 <mhroncok> jednorozec: continuously. yes, I upload it to fedorapople and I send emails once ina  while
16:57:50 <jednorozec> so you are ok if we move it into automation and push the report to fedorapeople automatically?
16:58:14 <jednorozec> or there was interesting ide to explore content resolver to also produce some nice reports
16:58:37 <mhroncok> if somebody wants to take care of it
16:59:07 <mhroncok> yet it's not really that good in exposing "this is impacted by orphaned/unwanted package" due to this dep chain
16:59:27 <mhroncok> the script needs to be rewritten using libsolv. using content resovleer for that seems like an overkill
16:59:39 <jednorozec> yep
17:00:03 <jednorozec> using libsolv was actually frist thing in my mind
17:00:12 <jednorozec> it will make the thing much simpler
17:00:15 <jednorozec> oh
17:00:20 <jednorozec> the time is up
17:00:24 <jednorozec> #endmeeting