#meeting-3:fedoraproject.org log

<@humaton:fedora.im>

17:00:38

!startmeeting RELENG (2024-01-16)

<@meetbot:fedora.im>

17:00:39

Meeting started at 2024-01-16 17:00:38 UTC

<@meetbot:fedora.im>

17:00:39

The Meeting name is 'RELENG (2024-01-16)'

<@humaton:fedora.im>

17:00:46

!meetingname releng

<@humaton:fedora.im>

17:00:51

!chair nirik sharkcz pbrobinson phsmoura dustymabe jednorozec

<@humaton:fedora.im>

17:00:53

!topic init process

<@nirik:matrix.scrye.com>

17:01:27

!hi

<@zodbot:fedora.im>

17:01:29

Kevin Fenzi (kevin) - he / him / his

<@humaton:fedora.im>

17:01:50

!hi

<@zodbot:fedora.im>

17:01:51

Tomáš Hrčka (humaton) - he / him / his

<@lenkaseg:fedora.im>

17:01:53

!hi

<@zodbot:fedora.im>

17:01:54

Lenka Segura (lenkaseg)

<@leo:fedora.im>

17:01:57

!hi

<@zodbot:fedora.im>

17:01:59

Leo Puvilland (leo) - he / him / his

<@leo:fedora.im>

17:02:06

just watching :)

<@patrikp:matrix.org>

17:02:19

Good evening to all.

<@pcreech:matrix.org>

17:02:30

!hello pcreech17

<@zodbot:fedora.im>

17:02:31

Patrick Creech (pcreech17)

<@humaton:fedora.im>

17:04:19

hello everybody, I hope you have a lovely day. Lets start with the agenda.

<@humaton:fedora.im>

17:04:22

.releng 11888

<@humaton:fedora.im>

17:04:26

!releng 11888

<@zodbot:fedora.im>

17:04:27

**releng #11888** (https://pagure.io/releng/issue/11888):**Fedora 40 Mass Rebuild Tracker** ● **Opened:** a day ago by jnsamyak ● **Last Updated:** 6 hours ago ● **Assignee:** Not Assigned

<@humaton:fedora.im>

17:04:52

So mass rebuild is scheduled to happen in about 24 hours

<@humaton:fedora.im>

17:05:15

changes to releng scripts and ansible should have appropriate PRs open

<@humaton:fedora.im>

17:05:27

But

<@humaton:fedora.im>

17:05:31

!fesco 3145

<@zodbot:fedora.im>

17:05:32

**fesco #3145** (https://pagure.io/fesco/issue/3145):**Delay mass rebuild for Fedora 40 until f40-build-side-81394 side-tag is merged.** ● **Opened:** 3 hours ago by codonell ● **Last Updated:** 22 minutes ago ● **Assignee:** Not Assigned

<@nirik:matrix.scrye.com>

17:06:26

i've updated the builders (but still f38)... and we finally got the last bits to enable dnf5 for builds this morning, so thats enabled.

<@nirik:matrix.scrye.com>

17:06:50

and also applied a cve fix to them. so hopefully they are all ready

<@jnsamyak:matrix.org>

17:06:58

!hello jnsamyak

<@zodbot:fedora.im>

17:06:59

Samyak Jain (jnsamyak) - he / him / his

<@jnsamyak:matrix.org>

17:07:03

sorry I'm kate

<@jnsamyak:matrix.org>

17:07:09

s/kate/late

<@nirik:matrix.scrye.com>

17:07:41

hey Samyak...

<@humaton:fedora.im>

17:07:48

hey jnsamyak

<@humaton:fedora.im>

17:08:04

SO I counted 4 +1 on the fesco ticket

<@nirik:matrix.scrye.com>

17:08:52

i'm not sure it really needed a vote. ;)

<@humaton:fedora.im>

17:08:55

if it gets through we will have to wait until the merge gets through

<@jnsamyak:matrix.org>

17:09:05

oh there's gonna be obe more +1 :)

<@jnsamyak:matrix.org>

17:10:09

oh there's gonna be one more +1 :)e

<@nirik:matrix.scrye.com>

17:10:24

sure

<@humaton:fedora.im>

17:12:01

ok so the bodhi update is not yet created

<@humaton:fedora.im>

17:12:16

should I go ahead and create it?

<@nirik:matrix.scrye.com>

17:12:20

<@nirik:matrix.scrye.com>

17:12:33

leave that to them when they are ready.

<@nirik:matrix.scrye.com>

17:12:35

imho

<@humaton:fedora.im>

17:13:16

ok so lets postpone the mass-rebuild for 1 day. Because this can easily endup merged in EU evening/night

<@humaton:fedora.im>

17:13:23

and that leads to errors

<@jnsamyak:matrix.org>

17:13:45

<@nirik:matrix.scrye.com>

17:14:24

yeah... agreed

<@jnsamyak:matrix.org>

17:14:32

also, I heard late and fast work is devil works so waiting a day will be best, I can update the epoch date in the prs if we all agree

<@humaton:fedora.im>

17:15:03

jnsamyak: Can you also send announcement or should I do that?

<@jnsamyak:matrix.org>

17:15:34

I can send one at devel-announce, in an hour post meeting :D

<@humaton:fedora.im>

17:16:14

oh I just checked and its 1045pm on your side

<@humaton:fedora.im>

17:16:16

ouch

<@humaton:fedora.im>

17:17:33

!agreed Mass rebuild for the Fedora Linux 40 will be done on 2024-01-18

<@nirik:matrix.scrye.com>

17:18:22

hopefully theres no more delays. ;)

<@jnsamyak:matrix.org>

17:18:56

🤞

<@humaton:fedora.im>

17:19:21

hope not but its much faster when we dont have to wait for aarch64 builds :)

<@nirik:matrix.scrye.com>

17:19:58

you mean armv7? ;)

<@humaton:fedora.im>

17:20:32

they should ahve named it leg64 and leave arm to be 32bit forever

<@humaton:fedora.im>

17:20:57

but also the new builders we got some time ago

<@nirik:matrix.scrye.com>

17:21:35

the slow ones these days seem to be ppc64le sadly

<@humaton:fedora.im>

17:22:22

Ok so lets move on.

<@humaton:fedora.im>

17:22:30

!releng 11870

<@zodbot:fedora.im>

17:22:31

**releng #11870** (https://pagure.io/releng/issue/11870):**Disable HTTP caches for repomd.xml files for Koji buildroots** ● **Opened:** a week ago by praiskup ● **Last Updated:** a week ago ● **Assignee:** Not Assigned

<@nirik:matrix.scrye.com>

17:23:15

saw it, too far down the pile for me to have loked into yet

<@nirik:matrix.scrye.com>

17:24:36

needs investigation. sadly theres lots of layers

<@humaton:fedora.im>

17:25:16

the proposed hotfix can be done by adjusting httpd config or?

<@nirik:matrix.scrye.com>

17:25:18

proxies / httpd -> haproxy -> varnish -> httpd

<@nirik:matrix.scrye.com>

17:25:31

unknown. needs digging

<@humaton:fedora.im>

17:25:38

I see

<@jnsamyak:matrix.org>

17:26:53

should we change tag to investigation, and medium trouble? I think it's tagged wrong

<@nirik:matrix.scrye.com>

17:28:31

if you like... didn't know we had an investigation tag... need to investigate that

<@humaton:fedora.im>

17:28:42

<@humaton:fedora.im>

17:28:50

I dont know about investigation tag

<@jnsamyak:matrix.org>

17:29:05

there is and i added it to the issue

<@humaton:fedora.im>

17:29:22

but what we agreed on previously was that medium trouble menas we need to investigate becouse we are not sure what is the root cause

<@humaton:fedora.im>

17:29:42

but what we agreed on previously was that medium trouble means we need to investigate because we are not sure what is the root cause

<@jnsamyak:matrix.org>

17:29:50

gotcha

<@humaton:fedora.im>

17:30:58

!releng 11800

<@zodbot:fedora.im>

17:30:58

**releng #11800** (https://pagure.io/releng/issue/11800):**Please add both obs-studio-libs and obs-studio-plugin-vkcapture to the multilib whitelist ** ● **Opened:** a month ago by gloriouseggroll ● **Last Updated:** a month ago ● **Assignee:** Not Assigned

<@humaton:fedora.im>

17:31:09

I think jnsamyak did some PRs for this

<@jnsamyak:matrix.org>

17:32:16

yeah, I can work on this, should be a quick addition to pungi-config multilib once I read go through the whole issue

<@humaton:fedora.im>

17:34:27

Didnt you already do that?

<@humaton:fedora.im>

17:34:32

or it was another package?

<@humaton:fedora.im>

17:35:04

here it is https://pagure.io/pungi-fedora/pull-request/1224

<@humaton:fedora.im>

17:35:39

you just need to do the same for that config in ansible repo in the bodhi role

<@humaton:fedora.im>

17:38:12

once the ansible PR is there I can merge both and redeploy

<@jnsamyak:matrix.org>

17:38:22

oh, I don't remember if it was the same package, but sure I can do that tomorrow

<@jnsamyak:matrix.org>

17:39:10

ok isee the pr and your comment there, sorry for missing it

<@humaton:fedora.im>

17:39:36

no worries :) I bet you I have bunch of abandoned PRs somewhere

<@nirik:matrix.scrye.com>

17:39:50

we all do ;)

<@humaton:fedora.im>

17:40:14

I just closed this one

<@humaton:fedora.im>

17:40:20

!releng 11826

<@zodbot:fedora.im>

17:40:20

**releng #11826** (https://pagure.io/releng/issue/11826):**Untretire EPEL branches for cxxopts** ● **Closed: Fixed with Explanation** 3 minutes ago by humaton ● **Opened:** a month ago by orion ● **Last Updated:** 2 minutes ago ● **Assignee:** Not Assigned

<@humaton:fedora.im>

17:40:32

this was a struggle with timings and automation :D

<@nirik:matrix.scrye.com>

17:40:35

yeah, how did you update pdc?

<@humaton:fedora.im>

17:40:46

I put it in the comment

<@humaton:fedora.im>

17:41:06

there is a script we have that uses client to update all slas for branch at once

<@nirik:matrix.scrye.com>

17:41:19

weird. I am pretry sure I used the same thing and it was telling me there was no epel8 branch.

<@nirik:matrix.scrye.com>

17:41:27

yeah, I know...

<@humaton:fedora.im>

17:41:33

weird

<@humaton:fedora.im>

17:41:46

but yeah I have seen some strange errors from PDC

<@nirik:matrix.scrye.com>

17:42:07

I sure hope we can kill it finally very soon.

<@humaton:fedora.im>

17:42:09

branch in api call in the web "UI" but nothing from client

<@nirik:matrix.scrye.com>

17:42:35

fun fact: pdc is the 2rd largest db we have on db01... (just behind hyperkitty)

<@humaton:fedora.im>

17:42:54

we will ceremonially burn the VM with release of 41

<@humaton:fedora.im>

17:43:07

composes composes

<@humaton:fedora.im>

17:43:15

there is a LOT of data in there

<@nirik:matrix.scrye.com>

17:43:29

yes, it used to store metadata on rpms too...

<@nirik:matrix.scrye.com>

17:43:34

anyhow, sidetrack

<@nirik:matrix.scrye.com>

17:44:05

although... how far are we from retiring pdc?

<@humaton:fedora.im>

17:44:35

from releng workflows I would say we have 50% covered

<@humaton:fedora.im>

17:45:07

but things will run in parallel for the 40 release cycle

<@nirik:matrix.scrye.com>

17:45:21

reminder: rhel7 goes eol on 2024-06-30... thats the hard deadline IMHO

<@humaton:fedora.im>

17:45:38

ok so we have to kill it at that date

<@jnsamyak:matrix.org>

17:46:40

sorry for interrupting folks, I drafted this email quickly, can someone read it through in case i missed any important details late night: https://hackmd.io/@2se5GAvIQGSWX9IGCuz2Tw/r1QukrNY6/edit

<@humaton:fedora.im>

17:47:27

go for it

<@jnsamyak:matrix.org>

17:47:35

shooot at us :D

<@nirik:matrix.scrye.com>

17:48:43

jnsamyak: looks fine to me, although I would change suppose to 'supposed' (which I know looks weird, but reads better). Or perhaps just change it to 'planned' ;)

<@nirik:matrix.scrye.com>

17:49:01

!releng 11823

<@zodbot:fedora.im>

17:49:02

**releng #11823** (https://pagure.io/releng/issue/11823):**please send openh264-2.3.1-4.fc40 to cisco** ● **Opened:** a month ago by kevin ● **Last Updated:** 2 days ago ● **Assignee:** patrikp

<@nirik:matrix.scrye.com>

17:49:14

whats the status here? no word from cisco?

<@humaton:fedora.im>

17:49:42

Weird thing happened here, they recieved all the parts of compose except for the first one

<@humaton:fedora.im>

17:49:59

So i resended it agan with response to theyr email before holidays

<@nirik:matrix.scrye.com>

17:50:02

:frowning

<@humaton:fedora.im>

17:50:13

and ofc it didnt go through the filter again

<@nirik:matrix.scrye.com>

17:50:14

ok, great. Hopefully they act on it soon.

<@humaton:fedora.im>

17:50:20

so they contacted me Yeasterday

<@nirik:matrix.scrye.com>

17:50:34

ugh. email is... not a good way to tranfer this.

<@humaton:fedora.im>

17:50:39

nope

<@humaton:fedora.im>

17:50:57

but should I just send them link to random internet place?

<@humaton:fedora.im>

17:51:08

or upload it to github?

<@nirik:matrix.scrye.com>

17:51:31

we have to be very very very careful there... we cannot distribute this.

<@humaton:fedora.im>

17:51:36

ofc that random internet place can be my server but it feels weird

<@nirik:matrix.scrye.com>

17:51:42

so it needs to be some way to only go to them and no one else.

<@humaton:fedora.im>

17:51:53

right

<@humaton:fedora.im>

17:51:59

so email it is

<@nirik:matrix.scrye.com>

17:52:16

I guess. anyhow, sounds like it's being worked on.

<@nirik:matrix.scrye.com>

17:52:25

might update the ticket with info so others know

<@humaton:fedora.im>

17:52:50

patrikp: can you split the composes into 15M chunks so it can get through filters?

<@nirik:matrix.scrye.com>

17:54:04

!releng 11787

<@zodbot:fedora.im>

17:54:05

**releng #11787** (https://pagure.io/releng/issue/11787):**create EPEL 10 package signing key** ● **Opened:** a month ago by carlwgeorge ● **Last Updated:** a month ago ● **Assignee:** humaton

<@nirik:matrix.scrye.com>

17:54:14

this is done right? or was there something left here?

<@humaton:fedora.im>

17:54:34

I made the keys added people but didnt put it anywhere

<@patrikp:matrix.org>

17:54:40

The f40 ones? (Or I guess all)

<@humaton:fedora.im>

17:54:49

patrikp: all

<@humaton:fedora.im>

17:55:06

and Ideally if you can use tars param --tape-size

<@humaton:fedora.im>

17:55:23

not split

<@nirik:matrix.scrye.com>

17:55:50

jednorozec: ok. We can add it to epel-release, but... epel-release-10... which doesn't exist yet, so I would say we should just close that ticket for now?

<@humaton:fedora.im>

17:56:20

yup closing

<@nirik:matrix.scrye.com>

17:57:22

oh, I wanted to talk about unretire's... some people seem to file those tickets before the re-review and we currently keep them open and tell them to let us know when it's done. I'd like to propose we instead close them and tell them to reopen them or file a new one when it's done.

<@nirik:matrix.scrye.com>

17:57:34

but also I guess we are closer to automating...

<@humaton:fedora.im>

17:57:58

That is what I did previously

<@humaton:fedora.im>

17:58:25

but then people started commenting on closed tickets or oppening new ones

<@humaton:fedora.im>

17:58:54

so yeah lets close it with "get back later"

<@nirik:matrix.scrye.com>

17:59:05

was reminded of this when someone opened like 10 unretires at the same time they just filed the re-reviews. ;(

<@jnsamyak:matrix.org>

17:59:36

but that makes sense, they can always reopen it; I'll keep that in mind as well, lessen the burden on tracker page as well

<@nirik:matrix.scrye.com>

18:00:01

ah... remembered another thing.

<@nirik:matrix.scrye.com>

18:00:10

!releng 11761

<@zodbot:fedora.im>

18:00:11

**releng #11761** (https://pagure.io/releng/issue/11761):**missing artifacts from f39-rc-1.5 (final release)** ● **Opened:** 2 months ago by kevin ● **Last Updated:** a month ago ● **Assignee:** kevin

<@jnsamyak:matrix.org>

18:00:27

!info mail has been sent to devel-announce for the postponement of mass rebuild

<@nirik:matrix.scrye.com>

18:01:14

This is done, but... the checksums in the torrents are not signed. ;( Can we sign those and remake those respin torrents?

<@nirik:matrix.scrye.com>

18:02:25

once thats done we cna close this.

<@jnsamyak:matrix.org>

18:03:31

ok I need to check this once, it was confusing because the torrent was created after we created the checksums for those respins

<@nirik:matrix.scrye.com>

18:03:32

oh and we are over time. ;(

<@jnsamyak:matrix.org>

18:03:44

do we need to so something more?

<@nirik:matrix.scrye.com>

18:03:47

yeah, somehow the signed one isn't the one in there.

<@jnsamyak:matrix.org>

18:05:03

weirddd, I can check but hopefully on thrusday/friday, or is it needed urgently nirik ?

<@nirik:matrix.scrye.com>

18:05:42

well, it's waited this long. ;) Please ping me if you need more info... we can look for sure

<@jnsamyak:matrix.org>

18:06:18

sure thing, need to check what happned there once, will reach ou thanksss :D

<@jnsamyak:matrix.org>

18:06:28

that's all from my end

<@jnsamyak:matrix.org>

18:08:19

hehe, I will be heading off to bed almost midnight, see ya all folks and thanks jednorozec for hosting us :D

<@jnsamyak:matrix.org>

18:08:24

jednorozec++

<@nirik:matrix.scrye.com>

18:08:35

night!

<@humaton:fedora.im>

18:08:38

!endmeeting