On Tue, Apr 21, 2020 at 09:56:58AM +0200, Clement Verna wrote:
On Sun, 19 Apr 2020 at 20:24, Pierre-Yves Chibon
<[1]pingou(a)pingoured.fr>
wrote:
.... snip
>
> Possibly, but it would mean it would need secrets to unlock the vault
> and we don't use valut for anything else currently. ;(
That almost applies to loopabull as well ;-)
(Well we do use it, but not much)
> I wonder how hard it would be to move loopabull into openshift... that
> might also help the secrets thing by allowing it to keep things in
> openshift secrets...
I wonder if there is a benefit from using loopabull at that point,
having a
simple fedora-messaging consumer running in openshift would work just as
well.
Yeah, also fedora-messaging makes it quite easy to setup the consumer and
there is not much boilerplate needed. So we could consolidate on 1 "bot"
consumer that listen to all the messages we are interested in and then
triggers different tasks.
I don't have a strong opinion, I was just thinking that maybe loopabull
would mean less maintenance on the longer term (ie focusing on witting the
tasks and tests for these tasks) but maybe the more we use it the more we
will have to touch loopabull upstream.
The code base isn't that big or difficult and I'm pretty sure Adam would give us
access if we need to, so this may not be a problem per say.
> Is loopabull moved to fedora-messaging? Or still fedmsg?
I have ported it to fedora-messaging, it only needed a couple of tweaks
from the
existing amqp plugin.
I may still need a release. I know we talked about it with Adam on IRC
not long
ago, but looking at github it looks like there was no recent release
made.
So the version deployed in the infra is still using fedmsg ?
Currently yes
Pierre