Hi Pierre,
Pierre-Yves Chibon wrote:
On Mon, Apr 09, 2018 at 01:28:02PM +0200, Dominik 'Rathann' Mierzejewski wrote:
On Monday, 09 April 2018 at 13:20, Pierre-Yves Chibon wrote: [...]
Finally the reason this has not been correctly announced is that we are still investigating the capacity of the system and the pipeline to check if it can handle the load of all the package in Fedora, but as said, the pipeline is already sending messages to the production bus, thus this behavior :(
The obvious thing to do is to stop sending these messages to the production bus until fedmsg_meta is updated and the update pushed to production. Why hasn't this been done already?
The basic answer is that I didn't have the hand to do it. So instead I released a new fedmsg-meta-fedora-infrastructure with support for these messages and deployed it in production. Messages about the allpackages pipeline should now make some sense :)
Is this deployed in production? I just pushed some changes to a fork on src.fedoraproject.org and recieved an email for each commit that looked like this:
Subject: fedmsg notification
Notification time stamped 2018-04-12 04:23:01 UTC
https://jenkins-continuous-infra.apps.ci.centos.org/blue/organizations/jenki...
If there's a change from what Michael reported, I can't see what it is. :)
I suspect this shouldn't be triggering at all for forks, which I believe two open infrastructure tickets cover:
https://pagure.io/fedora-infrastructure/issue/6575 https://pagure.io/fedora-infrastructure/issue/6791
For non-forks, I think it would also be ideal to not fire this off for each commit but rather once per push.
Thanks,