<@nirik:matrix.scrye.com>
16:00:04
!startmeeting riscv-sig
<@meetbot:fedora.im>
16:00:07
Meeting started at 2025-02-18 16:00:04 UTC
<@meetbot:fedora.im>
16:00:07
The Meeting name is 'riscv-sig'
<@nirik:matrix.scrye.com>
16:00:14
!meetingname riscv-sig
<@nirik:matrix.scrye.com>
16:00:14
!topic init process / agenda gathering
<@meetbot:fedora.im>
16:00:16
The Meeting Name is now riscv-sig
<@abologna:matrix.org>
16:00:31
sup
<@nirik:matrix.scrye.com>
16:00:45
morning everyone
<@rwmj:matrix.org>
16:00:49
hello
<@davidlt:matrix.org>
16:01:06
!hi
<@zodbot:fedora.im>
16:01:09
David Abdurachmanov (davidlt)
<@rwmj:matrix.org>
16:02:15
so we gonna start moving builders over to the new system (:-) half-ish joking)
<@nirik:matrix.scrye.com>
16:02:37
I have some infra updates... ;)
<@nirik:matrix.scrye.com>
16:02:42
I guess we can start with that? what other topics?
<@abologna:matrix.org>
16:02:49
let's start with that
<@nirik:matrix.scrye.com>
16:02:59
!topic Infra updates
<@nirik:matrix.scrye.com>
16:03:20
So, I have a hub, 2 x86 buildvm's and a compose host all setup now.
<@rwmj:matrix.org>
16:03:48
the build vms are running qemu, or they're literally x86-64 builders?
<@nirik:matrix.scrye.com>
16:03:51
Next thing I want to do is sort out the builders setup. My hope is to use keytabs...
<@nirik:matrix.scrye.com>
16:03:59
qemu/kvm
<@davidlt:matrix.org>
16:04:06
Some tasks don't require riscv64 native host.
<@nirik:matrix.scrye.com>
16:04:08
they are mostly for createrepos/newrepo or whatever
<@rwmj:matrix.org>
16:04:16
oh i see of course
<@abologna:matrix.org>
16:04:33
what's the compose host doing?
<@nhanlon:beeper.com>
16:04:44
!hi
<@zodbot:fedora.im>
16:04:46
Neil Hanlon (neil) - he / him / his
<@nirik:matrix.scrye.com>
16:04:48
There's a ticket about access. I'm gonna process that soon...
<@nhanlon:beeper.com>
16:04:57
sorry commops ran over 😃
<@nirik:matrix.scrye.com>
16:05:06
currently, nothing, but once we have something it could make images and repos, etc.
<@abologna:matrix.org>
16:05:31
I think we need the host to be riscv64 to build images
<@davidlt:matrix.org>
16:05:32
Pungi composes will not work (at least not without some work), but that's for later time.
<@davidlt:matrix.org>
16:05:51
Reminder: http://fedora.riscv.rocks/compose/
<@nirik:matrix.scrye.com>
16:06:02
yeah, it may not work out, but pungi calls things like runroot for those tasks normally.
<@nirik:matrix.scrye.com>
16:06:24
ie, compose would be the pungi host... gather repos, call koji to make things, etc.
<@jmontleon:fedora.im>
16:06:29
Do we have enough kojiadmins/sysadmins; i asked for kojiadmin only; I'm fine requesting both, or letting others who know more about koji handle it.
<@nirik:matrix.scrye.com>
16:07:02
I don't think we want too many, but happy to add people where they want to do work.
<@davidlt:matrix.org>
16:07:15
I would be happy to see more folks helping out.
<@nirik:matrix.scrye.com>
16:07:44
the sysadmin group will allow you ssh/sudo access. I can work with anyone who needs that to explain how it all works.
<@conan_kudo:matrix.org>
16:07:46
!hi
<@zodbot:fedora.im>
16:07:48
Neal Gompa (ngompa) - he / him / his
<@abologna:matrix.org>
16:08:18
sounds like something that only a very few people should have tbh
<@davidlt:matrix.org>
16:08:34
If everything works we are unlikely to use that access (I hope).
<@kashyapc:fedora.im>
16:08:37
Yeah; you don't want to hand out `sudo` access easily :)
<@nirik:matrix.scrye.com>
16:08:44
Also, we are making sure that fedora-messaging is going to be happy... we have a field to indicate primary/secondary messages from koji, we want to make that more configurable, but then hopefully we can get fedora-message bus messages for builds.
<@nirik:matrix.scrye.com>
16:08:58
agreed on all counts. ;)
<@nirik:matrix.scrye.com>
16:09:13
I just don't want to be a blocker for fixing something/doing something. ;)
<@kashyapc:fedora.im>
16:09:38
Yeah, a quick side on something you said above: on build VMs being "qemu/kvm" -- you just meant they're fully emulated guests, right?
<@nirik:matrix.scrye.com>
16:09:55
yep.
<@abologna:matrix.org>
16:10:03
IIUC all that was mentioned is x86_64, no?
<@kashyapc:fedora.im>
16:10:35
No, nirik mentioned there are some build VMs that are emulated RISC-V guests.
<@abologna:matrix.org>
16:11:00
that's not what I understood. nirik can you please confirm?
<@nirik:matrix.scrye.com>
16:11:05
no no
<@nirik:matrix.scrye.com>
16:11:10
they are x86_64
<@davidlt:matrix.org>
16:11:18
Those are x86_64 for generic non-arch tasks.
<@nirik:matrix.scrye.com>
16:11:23
they are only for tasks that can be done on any arch (newrepos, etc)
<@nirik:matrix.scrye.com>
16:11:38
actual builds have to be done on a riscv builder
<@davidlt:matrix.org>
16:11:46
We do the same in fedora.riscv.rocks setup (but we use aarch64 instead).
<@nirik:matrix.scrye.com>
16:11:56
(of which we have 0 until I sort out keytab stuff)
<@kashyapc:fedora.im>
16:12:16
Ah, noted.
<@nirik:matrix.scrye.com>
16:12:24
yeah, it was just easy to use x86_64 here because there is capacity on the same vmhost that has the hub
<@kashyapc:fedora.im>
16:13:09
Yeah, noted. I mixed up. (Didn't mean to put words in your mouth)
<@davidlt:matrix.org>
16:13:12
nirik: should I prep some sacrificial board for keytab experiments, or maybe provide instructions how to update your board to F41 + new firmware?
<@nirik:matrix.scrye.com>
16:13:17
So, my plan is to sort out builder auth (I can test with my builder), then we should figure next steps...
<@nirik:matrix.scrye.com>
16:13:27
yeah, either way
<@davidlt:matrix.org>
16:13:56
Ok, give me a day or so. I will could a disk image without initial setup enabled to make it easier. I tested a new firmware today, seems to work.
<@nirik:matrix.scrye.com>
16:14:48
ok, I'll poke at the keytab setup and see if I can get one. Do we want to go to the trouble of making hostnames match up? or just make generic named hosts?
<@kashyapc:fedora.im>
16:14:55
Yeah, noted. I mixed it up. (Didn't mean to put words in your mouth)
<@abologna:matrix.org>
16:15:00
I think we could take https://fedoraproject.org/wiki/Architectures/RISC-V/SpacemiT-Koji-Builder as a starting point and clean it up so that it applies to any freshly-provisioned F41 host
<@abologna:matrix.org>
16:15:50
seem reasonably straightforward, install a few packages, create a few configuration files, off you go
<@davidlt:matrix.org>
16:16:18
Well the hostnames are all fakes. They could be created generic created centrally, or we could keep what folks picked up.
<@abologna:matrix.org>
16:16:43
I wonder if we could have a simple `fedora-riscv-koji-builder` package that you can install and drags in all the software and contains all the generic configuration, so that you'd only need to add the host-specific client certificate
<@davidlt:matrix.org>
16:16:44
I don't really have preference here.
<@rwmj:matrix.org>
16:16:46
the hostnames are not actually used really, as the hub never needs to contact the builders?
<@nhanlon:beeper.com>
16:16:46
i think Jason already sorta did that, i guess :D
<@nhanlon:beeper.com>
16:16:46
<@nhanlon:beeper.com>
16:16:46
> I think we could take https://fedoraproject.org/wiki/Architectures/RISC-V/SpacemiT-Koji-Builder as a starting point and clean it up so that it applies to any freshly-provisioned F41 host
<@nhanlon:beeper.com>
16:16:46
> seem reasonably straightforward, install a few packages, create a few configuration files, off you go
<@nirik:matrix.scrye.com>
16:17:36
Richard Jones: they are used by the builder to auth to the hub... so they have to match the builder name in the hub, but the actual hostname can be anything, doesn't matter
<@davidlt:matrix.org>
16:17:45
The only thing that's important is that folks know which hosts are hosted by them.
<@rwmj:matrix.org>
16:17:50
yeah I mean the names have to match is all
<@nirik:matrix.scrye.com>
16:18:01
yeah, and we know which hosts are what if we need to direct specific jobs
<@davidlt:matrix.org>
16:18:11
Yeah, it's all fakes in general, just for auth.
<@kashyapc:fedora.im>
16:18:13
abologna: A bit like installing the `fedora-packager` RPM that sets up a Fedora packaging env?
<@nirik:matrix.scrye.com>
16:18:42
Yeah, that brings me to that... we should look at adding a koji profile to fedora-packager.
<@nirik:matrix.scrye.com>
16:18:47
(once things are working well)
<@abologna:matrix.org>
16:18:50
maybe. I mean, the instructions are straightforward enough as-is, I'm just thinking out loud whether it could be made even more trivial
<@davidlt:matrix.org>
16:19:09
Jason Montleon Neil Hanlon Richard Jones what do you think about the host names? Do you care keeping existing names?
<@kashyapc:fedora.im>
16:19:23
nirik: I can file a ticket to track; so that it's not buried in some meeting minutes :)
<@rwmj:matrix.org>
16:19:28
doesn't matter to me, whatever works
<@nirik:matrix.scrye.com>
16:19:48
it might be I can setup the hostnames in koji, but set their principal to something else generic? not sure...
<@davidlt:matrix.org>
16:20:04
It doesn't matter to me too, whatever makes it easier for nirik and Fedora Infra folks
<@nirik:matrix.scrye.com>
16:20:48
I'll try and look more into whats possible here...
<@nhanlon:beeper.com>
16:20:52
yeah i have zero preference in any direction
<@davidlt:matrix.org>
16:21:14
nirik: related question, do you plan to have ansible playbook or something to setup the boards?
<@nhanlon:beeper.com>
16:21:19
i know that the hostname displayed in koji can be different than the builder hostname, at least
<@nirik:matrix.scrye.com>
16:21:35
yep
<@davidlt:matrix.org>
16:21:38
Yeah, Koji hostnames are just for auth, otherwise they are fake names.
<@jmontleon:fedora.im>
16:21:50
I have no preference on hostnames, whatever works best
<@nirik:matrix.scrye.com>
16:22:06
fun fact: build hosts are just users with a bit set telling it they are a host
<@abologna:matrix.org>
16:22:16
davidlt: you mean to act as builders? is anything other than what's described in https://fedoraproject.org/wiki/Architectures/RISC-V/SpacemiT-Koji-Builder needed?
<@davidlt:matrix.org>
16:22:21
I think everyone who hosts the boards don't have preference 😄 nirik your call basically.
<@davidlt:matrix.org>
16:22:32
Yes, more is needed.
<@nhanlon:beeper.com>
16:22:43
the only stipulation I have is that we must use hexidecimal numbering /s
<@davidlt:matrix.org>
16:22:47
For example firewall configuration to block Internet access except to a few hosts.
<@abologna:matrix.org>
16:22:48
a lot more?
<@abologna:matrix.org>
16:23:17
maybe something like that already exists for x86_64 builders?
<@abologna:matrix.org>
16:23:31
and we can just tweak it so that it applies to riscv64?
<@abologna:matrix.org>
16:23:41
(x86_64 and other architectures like aarch64)
<@nirik:matrix.scrye.com>
16:24:07
I'm happy to poke at it and talk with folks out of meeting on it. I hope I can work on it this week... also, after I get things rolling here, I am probibly going to need to step back some to work on our datacenter move. I'll still be around, but hopefully things can be handed off more to you all for the actual work. ;)
<@nirik:matrix.scrye.com>
16:24:28
yeah, thats just a iptables template (and soon a nft one)
<@abologna:matrix.org>
16:24:54
is that stored anywhere public?
<@davidlt:matrix.org>
16:24:58
I am still surprised that it's iptables (after so many years).
<@davidlt:matrix.org>
16:25:02
https://pagure.io/fedora-infra/ansible
<@abologna:matrix.org>
16:25:03
in general, are the steps documented?
<@nirik:matrix.scrye.com>
16:25:18
yeah, we are moving to nft... but it's not been a high priority
<@davidlt:matrix.org>
16:25:32
https://pagure.io/fedora-infra/ansible/blob/main/f/roles/koji_builder
<@nirik:matrix.scrye.com>
16:25:50
role/base/template/iptables/ has the iptables templates.
<@abologna:matrix.org>
16:26:22
so I suppose we could open PRs that make that role applicable to a freshly-provisioned F41 riscv64 machine
<@nirik:matrix.scrye.com>
16:26:36
The firewall is a second line of defense... mock uses systemd-nspawn with network off too.
<@davidlt:matrix.org>
16:26:54
We might need some mods, for example, koji hub is different thus a different IP address should be listed on iptables.
<@abologna:matrix.org>
16:26:54
I assume most things would just apply as-is, but we'd need a different koji configuration and stuff like that
<@nirik:matrix.scrye.com>
16:27:03
I'm happy to make a template for people to use
<@abologna:matrix.org>
16:27:13
going off the existing configuration seems like the obvious way to go
<@davidlt:matrix.org>
16:27:21
Correct, systemd-nspawn by default runs with network disabled (in mock).
<@abologna:matrix.org>
16:27:22
instead of coming up with our own approach
<@abologna:matrix.org>
16:28:00
I could experiment with that with one of my boards
<@davidlt:matrix.org>
16:28:07
We also might want to reduce it (at least for now). While we can install "imagefactory", we probably don't need it.
<@nirik:matrix.scrye.com>
16:28:28
yeah. +1 to no IF...
<@jmontleon:fedora.im>
16:28:31
so are these rules (once we have them set up for our hub, etc.) applied in addition to firewalld, or do we disable firewalld and apply these rules?
<@kashyapc:fedora.im>
16:28:43
There's also `systemd-vmspawn`, if VMs are required ;-)
<@davidlt:matrix.org>
16:28:45
Technically all those packages are isntallable, but some of them aren't gonna be used. Do we still install it to avoid changing playbook?
<@abologna:matrix.org>
16:28:47
I have no previous experience with setting up koji builders, but I speak some Ansible
<@nirik:matrix.scrye.com>
16:28:59
we disable firewalld and just use those...
<@nirik:matrix.scrye.com>
16:29:17
hopefully we can remove imagefactory from primary playbooks soon too
<@davidlt:matrix.org>
16:29:48
Yeah, disable and just load iptables.
<@abologna:matrix.org>
16:29:54
IMO we should aim for minimal diff. if that results in unused packages being installed on the builder, that's fine
<@nirik:matrix.scrye.com>
16:30:48
there's only 2 deliverables using IF I think left.
<@nirik:matrix.scrye.com>
16:30:52
anyhow, yes
<@abologna:matrix.org>
16:31:06
ideally you'd flash the generic image, run the ansible part and be mostly done
<@abologna:matrix.org>
16:31:22
that's roughly how it works for other architectures right?
<@nirik:matrix.scrye.com>
16:31:48
yep. Well, many of them are vm's so ansible does the base install too first
<@nirik:matrix.scrye.com>
16:32:04
ok, I've got to go feed this cat thats meowing in my face... back in few.
<@abologna:matrix.org>
16:32:12
hopefully the provision and configuration steps are separated
<@davidlt:matrix.org>
16:32:18
Also we have different values (timeout, maxjobs, etc.)
<@davidlt:matrix.org>
16:32:32
So we need to make sure that's not default.
<@abologna:matrix.org>
16:32:39
makes sense
<@abologna:matrix.org>
16:32:51
we'll certainly need to diverge for some things
<@abologna:matrix.org>
16:33:06
let's just limit that to those we can't do without :)
<@abologna:matrix.org>
16:33:55
speaking of divergence. have we figured out how to deal with our git overlay yet? O:-)
<@davidlt:matrix.org>
16:35:15
No, what are your preference?
<@abologna:matrix.org>
16:36:10
personally I'd just create a group on GitLab if that's feasible. that makes it easy for people to contribute via MRs
<@abologna:matrix.org>
16:36:36
makes it possible to have shared ownership of repos too
<@abologna:matrix.org>
16:37:14
I said GitLab because I like it better, but I'd be okay with GitHub too if there's an overwhelming majority in favor of that solution
<@nhanlon:beeper.com>
16:38:06
<@nhanlon:beeper.com>
16:38:06
i demand cgit on a random ec2 instance w/ a public IP.
<@nhanlon:beeper.com>
16:38:06
kidding, of course. i'm amenable to anything
<@nhanlon:beeper.com>
16:38:15
especially gerrit-based solutions 😉
<@davidlt:matrix.org>
16:38:38
So I wanted to use whatever Fedora does, and keep it simple.
<@kashyapc:fedora.im>
16:39:05
Just so I understand, is the suggestion to move the currently hosted 'dist-git' sources on the overlay to there. (I.e. moving only those that differ from stock Fedora?)
<@davidlt:matrix.org>
16:39:08
I picked Gitea because it was easy to host, good performance (initially), and it supported the same APIs as GH.
<@abologna:matrix.org>
16:39:11
yeah that would have been my preference too, but apparently it's complicated to set up things so that we can have shared ownership? which IMO we do
<@davidlt:matrix.org>
16:39:43
If we do a 3rd party dist-git. We will need to add one more line in Koji Hub configuration and iptables template.
<@abologna:matrix.org>
16:40:04
yes. most builds will still happen from pristine Fedora sources, as is the case today
<@kashyapc:fedora.im>
16:40:05
FWIW, what davidlt says makes sense to me, and as an additional factor: Fedora is also going to move away Pagure to "Forjego" (if I got my spelling right).
<@abologna:matrix.org>
16:40:25
forgejo I think
<@abologna:matrix.org>
16:40:27
doesn't matter
<@nhanlon:beeper.com>
16:40:27
I will note that there is now a staging Fedora forgejo instance https://forgejo.apps.ocp.stg.fedoraproject.org/ though I am not clear what the SLA is on it (perhaps @nirik knows better)
<@davidlt:matrix.org>
16:40:29
Yes, but idea was to use forks and just use koji hub configuration to pick who is allowed to submit from a fork.
<@davidlt:matrix.org>
16:40:58
Forjego is based on Gitea (they forked it a year or more ago). I wouldn't mind that setup too.
<@kashyapc:fedora.im>
16:41:06
I guess he's cat-feeding; I'll give him a minute to catch his breath :)
<@nhanlon:beeper.com>
16:41:20
hehe, good call
<@davidlt:matrix.org>
16:41:28
Realistically I am mainly thinking: what makes nirik and Fedora Infra happy and less work for them.
<@abologna:matrix.org>
16:41:30
I think I remember discussing this. I feel that's intractable. we need to have a canonical overlay repo for each package
<@abologna:matrix.org>
16:42:12
if each build can potentially come out of a different fork it becomes impossible to track things
<@davidlt:matrix.org>
16:42:36
Not really as today you know which package is modified, and it's easy to look at SCM URL in Koji.
<@abologna:matrix.org>
16:42:41
well, not impossible. but way too much work
<@davidlt:matrix.org>
16:42:53
Yes, today all repos are in one place. You can find them without looking at Koji.
<@abologna:matrix.org>
16:43:31
a central point of synchronization is a must have IMO
<@davidlt:matrix.org>
16:44:07
But then we want it maintained by Fedora Infra?
<@davidlt:matrix.org>
16:44:24
If we are moving to Fedora Infra things on what we can use are limited.
<@abologna:matrix.org>
16:44:58
I don't think having it on Fedora infra is mandatory, unless nirik tells us that using GitLab or whatever is unacceptable
<@davidlt:matrix.org>
16:45:14
Don't get me wrong. I happy with a lot if it's not Pagure 😄
<@abologna:matrix.org>
16:45:24
of course over a long enough period of time the overlay will disappear and so we'll be on Fedora infra by definition ;)
<@nirik:matrix.scrye.com>
16:45:35
the sla on the staging forgejo is -10000 :) (ie, it's testing deployment/setup, and can be reinstalled anytime)
<@nirik:matrix.scrye.com>
16:45:55
we do have a gitlab.com/fedora community space.
<@nhanlon:beeper.com>
16:46:06
heh. noted ! thanks
<@davidlt:matrix.org>
16:46:11
I think, we need nirik input on what options exist for us.
<@nirik:matrix.scrye.com>
16:46:29
yeah, there's lots of options (kind of too many)
<@davidlt:matrix.org>
16:46:44
Sounds like Gitlab is an option 😄
<@abologna:matrix.org>
16:46:54
https://gitlab.com/fedora/sigs looks attractive
<@davidlt:matrix.org>
16:47:05
There is space for SIGs.
<@kashyapc:fedora.im>
16:47:07
FWIW, if Fedora's GitLab community space is an option, that sounds reasonable to me.
<@nirik:matrix.scrye.com>
16:47:08
yeah, I don't think waiting for forgejo is going to be a good idea... thats gonna take a while
<@nhanlon:beeper.com>
16:47:11
I proxy my vote to kevin to make the best decision that gives him time to do other things
<@davidlt:matrix.org>
16:47:11
https://gitlab.com/fedora/sigs
<@nirik:matrix.scrye.com>
16:47:16
yeah.
<@abologna:matrix.org>
16:47:21
davidlt: great minds ;)
<@abologna:matrix.org>
16:47:40
can we create our RISC-V SIG group and host the overlay there?
<@davidlt:matrix.org>
16:47:47
I am fine with Gitea, but the main call is nirik as I am not the one setting this up 😄
<@nirik:matrix.scrye.com>
16:47:58
I'm perfectly fine with that... I mean over time this will get smaller and smaller and by the time we go primary we won't need it anymore
<@davidlt:matrix.org>
16:48:10
I get to experience this from a bit different angle this time 😉
<@davidlt:matrix.org>
16:48:23
nirik: how do we use it?
<@davidlt:matrix.org>
16:48:37
Does it somehow connect to FAS, or something?
<@nirik:matrix.scrye.com>
16:49:03
The gitlab space? I can setup a space for sig there. Yeah, it's a group in fas that provides perms for the subproject
<@abologna:matrix.org>
16:49:39
nirik: so you'd set up riscv-koji so that it's possible to submit builds with the source coming from a repo under https://gitlab.com/fedora/sigs/risc-v/ right?
<@davidlt:matrix.org>
16:49:44
That sounds fine. So we need RISC-V SIG space, and get a few folks capable of creating/maintaining repos there.
<@nirik:matrix.scrye.com>
16:50:03
my only question would be if gitlab is using some load balancer or other anoying thing to allow in iptables/koji hub
<@davidlt:matrix.org>
16:50:11
We could start using it with the current Koji setup too. We don't really need to avoid.
<@davidlt:matrix.org>
16:50:13
We could start using it with the current Koji setup too. We don't really need to wait.
<@davidlt:matrix.org>
16:50:31
So that's the thing I was worried about.
<@abologna:matrix.org>
16:50:33
davidlt: that would have been my suggestion as well!
<@davidlt:matrix.org>
16:50:45
IP addresses, firewalls, etc.
<@davidlt:matrix.org>
16:50:59
There is probably a list of front-end IPs for Gitlab somehwere...
<@nirik:matrix.scrye.com>
16:51:20
yeah, needs investigating. I don't know off hand. we don't build anything directly from there currently.
<@nirik:matrix.scrye.com>
16:51:34
can someone add these to:
<@davidlt:matrix.org>
16:51:46
Found a list of IPs for Gitlab email servers.
<@nirik:matrix.scrye.com>
16:51:47
https://hackmd.io/FusNIhBJSzG4saHUHpQhJw
<@nhanlon:beeper.com>
16:51:50
fwiw if you've not associated your gitlab and fas, you can try this URL https://gitlab.com/groups/fedora/-/saml/sso
<@davidlt:matrix.org>
16:52:07
https://docs.gitlab.com/user/gitlab_com/
<@nirik:matrix.scrye.com>
16:52:29
if the ip situation is ok, I'm fine using the sigs space there.
<@abologna:matrix.org>
16:52:52
I think we can try setting it up and kicking off some build
<@rwmj:matrix.org>
16:52:52
hey everyone, 8 mins, plus some have to prepare for the next meeting
<@abologna:matrix.org>
16:53:07
worst case scenario we're just going to leave the space unused, no biggie
<@abologna:matrix.org>
16:53:49
we can start by creating a single repo and trying to trigger a build on riscv.rocks koji from there
<@abologna:matrix.org>
16:54:10
that'd be a good exercise
<@nirik:matrix.scrye.com>
16:54:11
well, we can look and figure out if it willwork first
<@nirik:matrix.scrye.com>
16:54:25
I can look.
<@nirik:matrix.scrye.com>
16:54:30
anything else before we close?
<@davidlt:matrix.org>
16:54:39
We will need to change builder configurations, but that's easy.
<@davidlt:matrix.org>
16:55:17
We need to discuss tags structure, NVR naming, and what/when we import RPMs. Probably not this meeting.
<@abologna:matrix.org>
16:55:36
I don't have anything. I'm really excited about the news that nirik brought to the meeting and about the future!
<@nirik:matrix.scrye.com>
16:55:37
yep.
<@nirik:matrix.scrye.com>
16:56:00
yeah, things are moving forward... should be exciting times ahead.
<@nhanlon:beeper.com>
16:56:04
indeed!! Thanks @nirik
<@abologna:matrix.org>
16:56:27
nirik++
<@zodbot:fedora.im>
16:56:28
Sorry, but Fedora Accounts user 'nirik' does not exist
<@davidlt:matrix.org>
16:56:31
We will break everything, and build it again. I think, this will be a 3rd Koji instance.
<@nirik:matrix.scrye.com>
16:56:39
ok, thanks everyone... lets continue in #riscv:fedoraproject.org
<@nhanlon:beeper.com>
16:56:45
my one thing is we're close to passing this to the magazine editors to polish and make an image for it -- https://etherpad.opendev.org/p/viZNqV0q0YaIGw-XVKic
<@nhanlon:beeper.com>
16:56:55
speak soon or forever hold your peace 😛
<@nirik:matrix.scrye.com>
16:57:07
oh nice. Thanks for working on that Neil Hanlon
<@abologna:matrix.org>
16:57:10
I'll give it another read just in case but it looked really, really good
<@nhanlon:beeper.com>
16:57:11
thanks to everyone who has helped!!
<@zodbot:fedora.im>
16:57:21
davidlt gave a cookie to neil. They now have 27 cookies, 8 of which were obtained in the Fedora 41 release cycle
<@kashyapc:fedora.im>
16:57:25
Neil Hanlon: FWIW, looks good to me, I gave it a full look.
<@abologna:matrix.org>
16:57:27
just the kind of promo I had in mind :)
<@zodbot:fedora.im>
16:57:33
abologna gave a cookie to neil. They now have 28 cookies, 9 of which were obtained in the Fedora 41 release cycle
<@nirik:matrix.scrye.com>
16:57:43
!endmeeting