16:00:18 #startmeeting RELENG (2022-02-22) 16:00:18 Meeting started Tue Feb 22 16:00:18 2022 UTC. 16:00:18 This meeting is logged and archived in a public location. 16:00:18 The chair is jednorozec. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 16:00:18 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:18 The meeting name has been set to 'releng_(2022-02-22)' 16:00:18 #meetingname releng 16:00:18 The meeting name has been set to 'releng' 16:00:18 #chair nirik sharkcz pbrobinson pingou phsmoura mboddu dustymabe ksinny jednorozec 16:00:18 Current chairs: dustymabe jednorozec ksinny mboddu nirik pbrobinson phsmoura pingou sharkcz 16:00:18 #topic init process 16:00:30 I am here 16:00:51 morning 16:03:10 #topic Beta freeze 16:03:37 everything except for the autosign playbook is done 16:04:19 Do we have SOP on how to run it? I remember nirik mentioning something about console and need to enable ssh on the machine 16:04:44 yeah, I can enable sshd if you are ready to run the playbook. 16:05:03 or I can run the playbook too if you like 16:06:29 nirik, changes are in ansible so anyone can run it 16:06:58 I don't see a sop tho, so we should make one. 16:07:05 I actually did it but than noticed that only stg is updating 16:07:55 * nirik runs it now 16:12:12 cool 16:12:24 lets move on 16:12:48 #topic Releng SOP cleanup 16:13:04 I have found few SOPs I have never seen or used 16:13:25 so want to ask you guys if they are usefull or can be removed/ merged to some other page 16:13:45 https://docs.pagure.org/releng/sop_package_announce_update.html 16:13:52 https://docs.pagure.org/releng/sop_multi_boot_media.html 16:13:56 sure, yeah, we should clean things up. I think also there was a though to move to docs.fedoraproject.org 16:14:10 yeah 16:14:21 but we should remove unused stuff before the move 16:14:26 both of those can be nuked. 16:14:38 https://docs.pagure.org/releng/sop_two_week_atomic.html 16:14:44 mailman3 doesn't support topics anyhow, and we haven't made a multiboot dvd in... a long time 16:14:56 nuke. No more atomic. :) 16:14:56 righ 16:15:04 I remember multiboot dvds 16:15:21 *some* time ago 16:15:41 times when computers had optical drives 16:15:58 yeah, the olden days for sure 16:16:03 https://docs.pagure.org/releng/sop_update_branch_last_known_good.html 16:16:13 this is cryptic ^^ 16:16:18 there is no information there 16:16:22 that I can decode 16:16:41 yeah, I don't recall the background on this one... nuke I would say... 16:17:22 https://docs.pagure.org/releng/sop_creating_live_translation_images.html 16:17:58 nuke, translations are now tracked in weblate... 16:18:48 that was all on this list for now 16:19:01 All can be nuked and the last known good was during the mash days 16:19:09 I will nuke the pages and gerenerate the pag 16:19:15 +1 16:19:28 I vaguely remember using it, but it was during my early days 16:19:53 mboddu, any tips on what is obviously not useful? https://docs.pagure.org/releng/sop.html 16:20:39 I am not sure about this - https://docs.pagure.org/releng/sop_clean_amis.html 16:20:53 I don't think we have ever added a content generator, have we? 16:21:21 nirik: We did, we used it in the initial stages of module development 16:21:28 jednorozec: https://docs.pagure.org/releng/sop_custom_spins.html can be nuked 16:21:44 I think we do still use the clean ami's script 16:21:57 Nuke - https://docs.pagure.org/releng/sop_package_announce_update.html 16:22:06 roles/fedimg/templates/clean-amis.py yeah... 16:23:40 Okay, other than the ami's doc, every other doc mentioned here can be nuked 16:25:27 * nirik has a topic/proposal on announcements and then one thing about fedora-repos whnever we get to open floor. 16:25:32 ack and nuked 16:26:55 lets look at the tracker 16:27:11 .releng 10656 16:27:12 jednorozec: Issue #10656: block_retired broken - releng - Pagure.io - https://pagure.io/releng/issue/10656 16:27:20 this has been here for a while 16:27:59 yeah, we need to fix it up. I haven't looked at it other than to see the traceback 16:28:07 I started looking into it today 16:28:21 there is issue with some koji params that we pass in one of the functions 16:28:33 sequence item 5: expected str instance, NoneType found 16:28:42 means 6th element of some list is None 16:28:48 the list is koji params list 16:28:53 I have to dig deeper 16:29:54 * nirik nods. 16:30:37 I have it on my todo list first thing tomorrow 16:30:40 That is on my list for today as well 16:30:46 :) 16:30:47 ok 16:31:28 mboddu, so if you are unable to fix it just ping me on releng channel with any progress so I can continue. 16:31:35 Sure 16:32:27 .releng 10656 16:32:28 jednorozec: Issue #10656: block_retired broken - releng - Pagure.io - https://pagure.io/releng/issue/10656 16:33:04 so my only question here is If we have a wat to get the BZ token from bitwarden into ansible 16:34:04 wat? 16:34:39 you posted the same ticket, but I think you are discussing the next one? 16:34:50 oh 16:35:00 yeah, we can put bugzilla api tokens in ansible-private for use with ansible. 16:35:23 .releng 10642 16:35:24 jednorozec: Issue #10642: The automated FTBFS reminders need to be disabled before the end of month - releng - Pagure.io - https://pagure.io/releng/issue/10642 16:35:46 nirik, yeah that was my question so I should put it into the ansible private, 16:36:14 than it should be simple change jsut to replace username&pass with api_tokn 16:36:15 yeah, there's a small block at the end of the vars.yml file there, fill that out and add it there. 16:36:53 that block just has a short 'what is this' where is it changed and how important it is and when it was last changed/created. 16:36:59 nirik: But the idea is get away from using ansible private and start using bitwarden, right? 16:37:21 not really... 16:37:33 ansible-private is for things that need to be referred to from ansible 16:37:41 bitwarden is manual things that admins have to enter. 16:38:02 There might be some way to leverage bitwarden for ansible, but thats down the road if at all. 16:38:15 Yeah, sorry 16:38:36 * mboddu is still thinking of awx 16:38:40 not at all, it's a reasonable thing to consider 16:38:56 I see there already is api key for bllockerbugz 16:39:10 * jednorozec will add the releng one 16:39:11 yeah, blockerbugs is already moved to using api key. ;) 16:39:32 I will move ftbfs weekly tomorow 16:40:17 ok, I think churchyard was working on that too... so might sync with him. 16:40:38 oh I see comment on the infra issue from him 4days ago 16:40:47 will ping him 16:41:07 one more issue and we can move to open floor 16:41:18 .releng 10601 16:41:19 jednorozec: Issue #10601: Update ejected from push because "Cannot find relevant tag for python-pyspike-0.7.0-1.fc34. None of ['f34-updates', 'f34-updates-pending'] " - releng - Pagure.io - https://pagure.io/releng/issue/10601 16:41:28 there is couple of tickets like this 16:41:32 .hi 16:41:33 Leo[m]1: Sorry, but user 'Leo [m] 1' does not exist 16:41:35 I think the issue is resolved 16:41:39 .hello leo 16:41:40 Leo[m]1: leo 'Leo Puvilland' 16:41:48 i dont usually do releng but hi :) 16:42:05 Leo[m]1, welcome 16:42:13 jednorozec: We fixed those issues, can you verify and close them? 16:42:22 yeah 16:42:22 Hello Leo[m]1 , welcome to releng meeting 16:42:30 will do 16:42:49 I fixed all those that I could find a while ago. 16:42:52 is there any way i could help? 16:43:18 Yeah I did some as well 16:43:34 so will go through them and close them if everything is ok 16:43:37 lets move to 16:43:39 #topic Open Floor 16:43:39 There's one wacky update that I am not sure how to fix aside from poking the db... 16:43:45 https://bodhi.fedoraproject.org/updates/FEDORA-2020-44d2abaa1b 16:44:09 I had a few things... 16:44:38 mboddu: are you going to be pushing updates now that we are in freeze? do you want me to do weekends? or ? 16:45:10 heh last freeze i did not push even one update 16:45:21 nirik: I can push the updates, dont worry about them 16:45:24 every time I woke up the ticket was closed again 16:45:29 Re weekends, I will let you know 16:45:40 sure, happy to help, just let me know. 16:45:49 Thanks nirik 16:45:58 I have a thing I want to do for both releng and infra... 16:46:08 And that one wacky update, poking db seems to be the only solution 16:46:26 There's 'standard' announcements we make as part of the release cycle... 'freezes' and branching and such... 16:47:06 I am working on putting them into a doc in releng repo though, if there is another repo that you would suggest, I am happy to push it there 16:47:11 I'd like to check them all into git and add a link in the footer saying "This announcement is in XYZ repo, PR's and improvments welcome" 16:47:22 lets put them in git. 16:47:37 easier for people to PR and they are just text docs 16:47:53 +1 16:48:09 "lets put them in git" you mean another repo? 16:48:18 no, just releng git... 16:48:26 an announcements/ dir next to scripts? 16:48:35 Yup, thats what I am thinking 16:48:39 we can have some template=script so it spits correct versions in announcements automatically 16:48:58 ^ I am not sure about this 16:49:10 yeah we could at some point, but first step is to get them in git and use those... 16:49:18 Yup 16:49:23 then we can look at sending from 'releng' user or whatever 16:49:39 Oh I get it now, have a script to generate the announcement, thats nice 16:49:44 * mboddu needs more tea 16:50:14 yeah, announcment templates. and a script to generate the actualt one you need for f36 freeze for example 16:50:32 The last thing I had was fedora-repos... can someone add the f38 key and update all releases? There was a request for f34 to get f37 key at least, but IMHO it would be good to push f38 out too so we are ahead... 16:51:01 +1 16:51:05 I never did it 16:51:11 so would like to do it 16:51:17 Sure, jednorozec I will help you to do it 16:51:36 https://bugzilla.redhat.com/show_bug.cgi?id=2055220 16:51:44 oh, reminds me... 16:52:16 There's some fedora-repos PRs 16:52:18 https://src.fedoraproject.org/rpms/fedora-repos/pull-request/114 16:52:40 I really don't like this, but haven't had time to articulate to the submitter. 16:53:42 But feel free to comment if you have opinions. :) Anyhow, thats it I think from me. 16:54:04 I have some opinions, I need to respond to the PR 16:55:13 oh so it is the 2s load time that is a problem 16:55:33 now I get why they want it 16:55:50 seems like they could just import current and current-1 and current-2 16:56:06 oh, and container builds seem broken again. 16:59:26 containers oh containers 17:00:27 ok 17:00:32 let me close the meeting 17:00:36 #endmeeting