16:03:22 #startmeeting RELENG (2023-05-16) 16:03:22 Meeting started Tue May 16 16:03:22 2023 UTC. 16:03:22 This meeting is logged and archived in a public location. 16:03:22 The chair is nirik. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 16:03:22 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:03:22 The meeting name has been set to 'releng_(2023-05-16)' 16:03:23 0/ 16:03:23 #meetingname releng 16:03:23 The meeting name has been set to 'releng' 16:03:23 #chair nirik sharkcz pbrobinson phsmoura dustymabe jednorozec 16:03:23 Current chairs: dustymabe jednorozec nirik pbrobinson phsmoura sharkcz 16:03:23 #topic init process 16:04:10 Good evening. 👋 16:04:32 hello everyone 16:04:48 hopefully jednorozec will be along soon... 16:05:41 There is a possibility that he might currently be in transit. 16:05:59 yeah, he was... 16:06:02 hello 16:06:15 hey 16:06:18 train delays the usuall stuff 16:06:30 Hey! 16:06:42 hey. big crowd today! 16:07:01 I also cooked ramen for dinner 16:07:25 and meeting ;) 16:07:34 nice 16:07:58 🍲 16:08:23 anyhow, I had a few items... 16:09:13 #topic toolbx container 16:09:20 .releng 11399 16:09:21 nirik: Issue #11399: Add container/fedora-toolbox as a Fedora Linux 39 release-blocking deliverable - releng - Pagure.io - https://pagure.io/releng/issue/11399 16:09:49 jednorozec: were you gonna look into adding this to pungi/composes? 16:10:28 I think we pretty much need to to make it release blocking 16:10:37 yes we do 16:12:16 sorry all the animals are welcomiing me home 16:12:19 anyhow, we should try and get it sooner rather than later... in case we need to adjust pungi or something. 16:12:24 cats jumping on keyboard dog howling 16:12:27 it's nice to be loved. ;) 16:12:56 so I was going to work on the toolbox this week 16:13:14 I was typing slow because a cat had immoblized one arm... she decided to go lay in the sun tho now. ;) 16:13:20 cool. 16:13:26 #topic builder upgrades 16:13:44 I was just going to do builder upgrades in the outage tomorrow... unless I run out of time or something. 16:13:59 is that ok? or did you want to do a seperate outage for them later and you do them? 16:14:14 I dont mind 16:14:32 maybe I can stick around and do the upgrades during outage? 16:15:08 oh its 21 UTC so probably not :D 16:15:20 also related. I just disabled all the a32 build vm's... and we can repave those bvmhost-a64's as buildhw builders at some point. 16:15:20 sure, if you like! I secheduled it for 21UTC tho... so not sure if thats too late for you... 16:15:25 yeah. ;( 16:15:35 we will see 16:15:53 finally armhf 16:16:29 I think I had one other thing, but now I can't remember what it was. 16:17:12 oh, that was it! 16:17:18 #topic autosign 16:17:44 I'm going to try and migrate from autosign01 (rhel8, old hardware) to autosign02 (rhel9, new hardware)... probibly next week? 16:18:11 hopefully it will go smoothly... just a heads up. Should just be installing stuff and configuring and it should all just work. 16:18:19 (famous last words) 16:18:54 Thats all I had off hand... you have any topics jednorozec? or should we triage some old tickets? or just open floor it? 16:19:03 .releng 11422 16:19:04 jednorozec: Issue #11422: Please send openh264-2.3.1-1.el9 to Cisco - releng - Pagure.io - https://pagure.io/releng/issue/11422 16:19:13 considering 16:19:25 .releng 11425 16:19:26 jednorozec: Issue #11425: openh264 is not a multilib repo - releng - Pagure.io - https://pagure.io/releng/issue/11425 16:19:47 should I make compose for all active releases with multilib and send to cisco? 16:19:51 including that el9 16:20:11 yeah, I think so... if we can figure out how to do it. 16:20:16 right 16:20:17 so 16:20:26 I looked briefly on dist-repos 16:20:51 can we make the results not accessible? so nobody randomly downloads it from us? 16:21:30 it would be super simplified over the current odcs stuff 16:21:34 we already block h264 rpms on kojipkgs... but we will want to make sure it's working. 16:21:55 that also means you have to copy from the nfs mount, can't download them and upload them somewhere. 16:22:01 (but thats the case with odcs too) 16:22:13 odcs does do multilib, it's just a config option. 16:22:23 yup it does 16:22:26 so, perhaps we should just do that for now and look at moving to dist-repos after? 16:22:33 ack 16:22:50 I can show guys in the morning meeting how to do openh264 16:23:45 soo that is also all I have for today 16:23:56 quick one it seems 16:24:00 sounds good. So just add mulitlib there, do new repos, send em, and then we can test some dist-repos and see if we can switch 16:24:22 yup that is the plan 16:24:43 I have to create 2 tickets one for tracking this and one for tracking the odcs backends unification 16:24:47 sounds good. 16:25:09 #topic open floor 16:25:17 I have a question. 16:25:24 go on 16:25:31 What are the differences between a compose and a system image? 16:25:55 in context of pungi? 16:26:58 * nirik isn't sure what a system image is... a compose is a full build of all artifiacts and deliverables. ;) 16:27:11 https://en.wikipedia.org/wiki/System_image 16:27:31 "In computing, a system image is a serialized copy of the entire state of a computer system stored in some non-volatile form such as a file." 16:27:31 Are they synonyms then? 16:27:40 nope 16:27:54 in our context and context of that wiki page 16:28:07 only system images we produce are docker images and qcow virtual images 16:28:13 and they are product of compose 16:28:35 huh, ok. The images we build are just bootable images, the state after the boot is not something we control or save... 16:29:36 just that is not true for the docker 16:29:51 thatis kind of that thing from wiki definition 16:29:55 yeah, I guess so... 16:30:19 academic debate :D 16:30:49 jednorozec: So composes come first and system images are created from them? 16:30:58 patrikp[m], nope 16:31:06 compose is a process 16:31:11 and also a result 16:31:20 that is all the artifacts we have 16:31:38 so when we run a compose it can be full one with all the deliverables 16:31:57 or one that composes just repositories like updates and updates-testing 16:32:20 one we fixed this morning, was a compose but no images were created 16:33:04 and compose is also what nirik wrote, full build of all artifiacts and deliverables 16:33:30 jednorozec: Ah. So when it is used as a noun, it's a compose (process). 16:34:49 More basic question then... 16:35:22 "isn't sure what a system image..." <- > * <@nirik:matrix.scrye.com> isn't sure what a system image is... a compose is a full build of all artifiacts and deliverables. ;) 16:35:23 What are the definitions of "artifacts" and "deliverables"? 16:35:35 so 16:35:59 https://docs.fedoraproject.org/en-US/releases/f38/blocking/ 16:36:10 those are release blocking deliverables 16:37:26 deliverables I think of as "Things we deliver to the customer(s)" and those consist of a number of different artifacts (a deliverable might have a number of different artifacts, like a image and a install tree, etc) 16:38:24 also rpm build can be considered artifacts, basicaly anything that is produced by the compose process or build system 16:39:28 does it make sense? 16:39:33 yep just on more clarified definitions I refer these from my notes: (correct me if it is wrong) 16:39:33 Artifacts are tangible or intangible items that are created, modified, or used during the project lifecycle. These items serve as documentation, builds, etc. 16:39:33 Deliverables, on the other hand, are specific tangible or intangible outputs that are produced at the completion of a project or at specific project milestones. 16:40:27 sounds reasonable, yeah 16:40:54 yup 16:42:56 Would a system image fall under the (broader) category of a "deliverable" then? If we go by the definition of system image that I quoted earlier. 16:44:20 in my opinion yes 16:44:54 if we had some but we dont 16:45:25 except for the docker but that is a different story because its expecting some state on the host 16:45:52 just the first sentese from the definition is something we just do not do 16:46:40 "isn't sure what a system image..." <- > * <@nirik:matrix.scrye.com> isn't sure what a system image is... a compose is a full build of all artifiacts and deliverables. ;) 16:46:40 And now that we established what deliverables and artifacts are... What is the definition of a "full build"? 16:47:35 well, a build of all the artifacts and deliverables for whatever milestone/release 16:47:50 not just some of them, but all you require 16:48:41 Alright. I think I am starting to get it. Thank you very much. 16:49:12 * jnsamyak liked the discussion :D 16:49:23 one quick thing 16:49:31 just an update on f36 EOL, the releng koji_signed_cleanup script is still running, hopefully is on some p* named package so might get over soon, I'm watching over it; Once done, I'll shoot an email; 16:49:31 For that I modified a template a little bit (the tense), the draft is here: https://paste.centos.org/view/837232e7 16:49:31 If it is okay, I'll post this to email list (+1 are appreciated!) 16:51:55 I'd say "will be available" instead of "were available"... and 'f37 will continue to' instead of 'f37 continued' 16:52:08 yep 16:52:20 will do that 16:52:20 and 'Fedora Linux releases are documented' instead of 'Fedora Linux releases was documented' 16:52:35 but looks good otherwise. +1 16:53:53 I only speak czenglish so no input on this from my side 16:54:10 ha. 16:54:36 I would say Fedora GNU/Linux... But that is a topic for another time, another open floor. :-) 16:55:15 ok, lets close out then... thanks everyone for coming. :) Much nicer when there's more people to discuss things. 16:55:21 #endmeeting