16:00:15 #startmeeting RELENG (2022-02-01) 16:00:15 Meeting started Tue Feb 1 16:00:15 2022 UTC. 16:00:15 This meeting is logged and archived in a public location. 16:00:15 The chair is jednorozec. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 16:00:15 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:15 The meeting name has been set to 'releng_(2022-02-01)' 16:00:15 #meetingname releng 16:00:15 The meeting name has been set to 'releng' 16:00:15 #chair nirik sharkcz pbrobinson pingou phsmoura mboddu dustymabe ksinny jednorozec 16:00:15 #topic init process 16:00:15 Current chairs: dustymabe jednorozec ksinny mboddu nirik pbrobinson phsmoura pingou sharkcz 16:00:59 .hi 16:00:59 phsmoura: phsmoura 'Pedro Moura' 16:01:02 hi, guys 16:01:20 morning 16:02:16 hello 16:03:34 so lets talk mass rebuild 16:03:38 .releng 10538 16:03:40 jednorozec: Issue #10538: Fedora 36 Mass Rebuild - releng - Pagure.io - https://pagure.io/releng/issue/10538 16:04:07 I think we did not have any issues with the rebuild it self 16:04:47 but thanks to outdated docs I re-signed the result 16:05:07 yeah, it happens, no worries. 16:05:38 yup, the doc is now updated 16:05:47 * jednorozec re-cheks 16:06:10 yup, updated https://docs.pagure.org/releng/sop_mass_rebuild_packages.html#post-mass-rebuild-tasks 16:07:15 there is one more thing, I started to rewrite find_failures.py we use to generate the html report 16:07:26 to include failed source builds 16:08:41 any help will be appreciated, testing changes with bugzilla api can take some time 16:09:26 closing the mass rebuild tracking ticket 16:09:32 cool. thats been a long time bug. ;( 16:09:49 heh, it is actually not that easy to get that information 16:10:08 I am in meeting with asaleh to go over bodhi issues 16:10:11 jednorozec: do yyou think is a good ideia for me to help with that. to learn more about parts of the infra like this bugzilla api? 16:10:41 nah I mean koji api 16:11:38 phsmoura, so here you go https://pagure.io/releng/blob/main/f/scripts/find_failures.py 16:12:21 mboddu: the bodhi issues were likely koji 16:13:16 this scripts get list of failed builds, we want it to include also failed source rpm builds. I can give you some hints after the meeting 16:14:29 closed the mass rebuild issue 16:14:40 nirik, can you elaborate on what happened? 16:15:40 The tagging issue we hit with 1.27.x... https://pagure.io/koji/issue/3220 16:16:02 basically you would ask something to be tagged, it would be tagged, but then return failed as already tagged also 16:16:55 so bodhi would say 'I'm doing an updates push, I want you to tag these 20 builds' and 5 of them would return failed, so the compose would fail. But the tagging worked, so bodhi was confused as to the state. 16:18:04 right 16:18:08 there is PR already 16:19:16 cool. I can build that today and re-upgrade the hubs 16:20:12 so I have a question 16:20:18 about side tags 16:20:26 there are tickets like 16:20:31 .releng 10607 16:20:34 jednorozec: Issue #10607: f36-java17 named sidetag - releng - Pagure.io - https://pagure.io/releng/issue/10607 16:20:54 my question is about our SOP and signing 16:21:14 should I really add the side tags to ansible for 3 weeks and remove them later? 16:21:44 or merge the sidetag to -pending at the end of use 16:22:09 well, if it's really large it could be disruptive... 16:22:21 but I'm not sure how much it matters. 16:22:36 nirik: We are looking at the ejections issue (before the recent bodhi update) 16:22:38 500 builds isn't too many really 16:22:48 that was my thinging 16:22:52 the gnome one was like 200 16:23:14 mboddu: well, I looked at it a while back and tracked it to the koji tagging issue... but if you find some other cause, great. 16:26:10 I will go ahead and update the side tag SOP so the person reading it will consider if it makes sense to add it into auto-sign or sign upon merging 16:27:17 sounds reasonable 16:29:27 moving on 16:29:30 .releng 10394 16:29:31 jednorozec: Issue #10394: Unretire script doesn't properly unretire packages - releng - Pagure.io - https://pagure.io/releng/issue/10394 16:29:41 phsmoura, this could be also interesting 16:31:27 cool, im reading it 16:33:07 phsmoura, one of our community members already implemented the new API endpoint https://pagure.io/fork/mattia/pagure-dist-git/commits/unretire 16:34:03 just get the tests running and open PR, next new unretirement scripts need to be updated 16:36:39 ok, think I got. thanks 16:37:18 do you have any updates? nirik mboddu sharkcz phsmoura 16:37:40 updates on? anything? 16:38:13 Sorry, I am still with Adam and I dont have any updates 16:38:30 nirik, anything releng related? 16:38:55 If not I propose to go through our old tickets: https://pagure.io/releng/issues?status=Open&order_key=date_created&order=asc 16:39:02 well, rawhide has been failing... for pretty wacky reasons. ;) 16:39:04 Nothing from me, just couple of questions regarding packaging tools, but I still want to try some things before asking 16:39:18 nirik, this brings me 16:39:34 nirik: Yeah, it is the tagging issue but bodhi is not tagging them properly, which is what we are trying to figure it out 16:39:44 I was to update the compose report thingy to close the tickets when opening new ones 16:39:45 I worked around the first issue, now there is a new one, but I suspect it's due to the hub being different koji version from the builders. 16:40:21 mboddu: well, note with this bug koji says "failed!" but actually it has already been tagged... so I can see why that would confuse bodhi. 16:40:37 jednorozec: awesome 16:40:53 that tracker is a mess 16:41:26 my question is how many old composes should stay as opened tickets? 16:41:30 last 10? 16:41:36 nirik: Yeah, which is what we are trying to figure it out 16:42:12 10 sounds reasonable... or we could do 14 days which is what we keep for composes I think? 16:42:21 oh 16:42:24 that make sense 16:42:45 It will close all older 14 days 16:42:53 than the last ticket it opened 16:45:08 +1 16:49:13 from the department of old tickets I bring you 16:49:16 .releng 7337 16:49:17 jednorozec: Issue #7337: Emit fedmsg when candidate composes are synced to stage - releng - Pagure.io - https://pagure.io/releng/issue/7337 16:51:44 so the sync is a manual step 16:52:07 Yeah, its a rsync 16:53:02 SO we need to emit the message by some script 16:53:19 Yup 16:55:26 what if we create ansible role and point people to the message that is emitted by running ansible? 16:56:44 well, it shouldn't be too hard to make a small script... rawhide/branched composes do this... 16:56:52 Yeah, I prefer ^ 16:56:58 or perhaps we could even reuse that... not sure 16:59:07 rawhide/branched used https://pagure.io/releng/blob/main/f/scripts/fedora-messaging.sh as https://pagure.io/pungi-fedora/blob/main/f/nightly.sh#_41-47 16:59:15 We could just make use of it 17:01:34 ok 17:03:45 I wrote comment there with the links 17:03:52 and the time is up 17:03:56 #endmeeting