15:02:08 #startmeeting RELENG (2023-04-18) 15:02:08 Meeting started Tue Apr 18 15:02:08 2023 UTC. 15:02:08 This meeting is logged and archived in a public location. 15:02:08 The chair is jednorozec. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 15:02:08 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:02:08 The meeting name has been set to 'releng_(2023-04-18)' 15:02:08 #meetingname releng 15:02:08 The meeting name has been set to 'releng' 15:02:09 #chair nirik sharkcz pbrobinson phsmoura dustymabe jednorozec 15:02:09 #topic init process 15:02:09 Current chairs: dustymabe jednorozec nirik pbrobinson phsmoura sharkcz 15:02:19 nirik saves the day 15:02:25 morning again 15:03:11 beautiful sunny afternoon on this hill 15:03:21 sooo 15:03:25 we had snow this morning. 15:03:44 haha I got mine over Easter 15:04:22 getting to dangerously comfortable 20C for like 1h or so 15:04:31 ha 15:04:37 so I had a few small things... 15:05:09 go on 15:05:12 #info fedora 38 is out! get it today! 15:06:02 I talked with mboddu the other day, and he cannot make this meeting time (he has a conflict), but he would be happy to meet with us some other time to go over old tickets. I think he might have knowledge of some of the old ones... 15:06:18 so perhaps we could schedule a special meeting next week sometime or something? 15:06:27 sure 15:06:29 just start with the oldest releng tickets and work back from there. 15:06:50 nirik: +1, some other time should work 15:07:10 could you look at scheduling that? or would you like me to? 15:07:12 I am ok with any time before 21:00 CET after that I cannot guarantee my cognitive abilities 15:07:32 I'm not sure mohan's availability, but we can check calendars, etc. 15:07:56 I can do that 15:08:03 let me open calendar right now 15:08:24 cool. Doesn't have to be immediate, but whenever. ;) 15:09:00 looks like we all have a slot in 1h and 22mins next week 15:09:02 Next item: now that f38 is out we need to update builders. I was thinking we had enough other stuff going on that I would just update them instead of reinstalling. Much easier. 15:09:05 is that ok? mboddu nirik ? 15:09:46 cool. Send an invite? Others might want to join too... we could meet on irc/matrix? 15:10:01 nirik, like running dnf system-upgrade on them? 15:10:28 yeah, or just dnf --releasever 38 distro-sync, reboot... should work fine. 15:10:35 right 15:10:38 reinstalling them is a lot more pain. 15:10:45 but sometimes good. 15:11:06 I prefer place with meeting bot so we can have history and potential notes for future reference 15:11:41 sounds good to me. ;) 15:11:48 might mention it on the releng list or something too. 15:15:04 hmm calendar hell 15:16:31 yeah, happens. ;) 15:16:40 Lets see... I don't think I had much else... 15:17:10 ok now the correct invite is in place 15:17:14 pardon the spam 15:17:34 sigh, now I have cats growling at each other for breakfast. 15:17:56 so 15:18:16 I think I have something but it got lost in the torrents and IoT stream 15:18:26 ok, I closed the troublemaker into a room. ;) 15:18:45 oh there was this in the past week https://pagure.io/releng/failed-composes/issue/4870 15:19:16 samyak7, fixed it today and I shoved him how to restart the compose cron job 15:19:35 spell check fail 15:19:55 ah yeah. Can we fix the rawhide branch to be set for f39 down the road? 15:20:22 in the main branch? 15:20:36 jednorozec: +1 15:20:39 yes 15:21:07 sure thing 15:21:10 we should also look at abstracting out things in that repo... ie, have a fedora-common.conf and just really small container/nightly/etc ones. 15:21:22 yes 15:24:10 ok i have brain fog, nothing in this head 15:24:20 * jednorozec scrolls the tracker 15:24:45 oh right 15:24:54 how do we create tokens in pdc? 15:25:11 not easily. 15:25:21 anton opened ticket with fedora-infra about it some time ago 15:25:38 we need one to do PDC changes for unretirement automations 15:25:47 so in the past, you could login with SAML2, then scrape the token out of your web browser.. and we could make it an admin one in the database. 15:25:59 but... SAML2 auth is currently busted. 15:26:04 (on pdc) 15:26:20 hum 15:26:40 I guess we could try and just generate and insert a new one into the db directly? 15:26:49 there is probably no such nice tool as in pagure to create such tokens right? 15:27:05 nope. :( 15:27:13 we could push for finally retiring pdc... ;) 15:27:31 ORM based project can go crazy when one just write things into DB 15:27:55 yeah. I'm not crazy about the idea. 15:28:07 now we have manpower to do the ops stuff, I am going to snad by my word and retire that thing 15:28:14 we could also try and fix SAML2 auth. I cc'ed Aurélien B in the ticket about it. 15:28:21 ceremonially burn it by stream of bytes 15:28:41 ha. I'll roast a marshmellow on it. :) 15:29:03 we did a ARC on retiring it a while back right? ie, we identified all the parts we need to move to somewhere else? 15:30:03 yes 15:30:23 there is a doc in the shared space 15:30:23 anyhow, we should do it carefully. Make sure everything that uses it is handled and it's getting no traffic and then shut it off, then nuke it after a week or two 15:30:37 great. :) 15:30:53 oh for sure, we will find some wires hiding somewhere in the apps 15:31:35 just changing the src.fp.o git hook to not query it will cut it's traffic down a ton. 15:31:59 I was thinking about that 15:32:39 what if we used protected branch to store a file with EOL FOR each branch? 15:32:43 we may need to try and drive adamw's release thing forward (so we can have a single place to query release versions, eols, etc) 15:33:13 ie, it will be writable only by group releng so we will be able to modify it(toddlers) and users dont 15:33:16 well, if we were doing that we could just make it a config option? 15:33:40 hmm 15:33:42 but yeah, stream branches need dealt with. 15:33:57 anyhow, I suspect this was covered in the arc thing... 15:34:07 yeah something are there 15:34:55 we will revisit it any how, meanwhile I will share the one token I used through ansible-private just to the deployment 15:35:36 we may be able to fix SAML2... not sure. 15:35:37 so we can move the unretiement to selfservice with newcommers watching over it 15:35:55 ok. Hopefully it handles all the cases well. ;) 15:36:07 probably not :D 15:36:16 but it should not cause any fires 15:36:27 maybe, probably 15:36:31 we will see :D 15:36:40 I'm happy to look at it before we roll it out... 15:37:08 michal already took over the code review 15:37:25 But there is a PR in the toddlers 15:38:04 so that was all I had 15:39:18 cool. Shall we meet next week? or just have the ticket backlog meeting? 15:41:56 lets keep it opened and see if somethings comes up during the week 15:42:26 #endmeeting