Description:
The hackingand.coffee ( https://www.archlinux.org/mirrors/hackingand.coffee/ ) mirrors are migrated to a load balancer that handles mirror downtime on our end. This helps mitigate mirror maintenance issues by removing it from the pool on our end, and not having to rely on the mirror system externally. We used to do this with a virtual balancer back when we first setup multiple hardware nodes, but that created issues. Now that we have an extra hardware balancer at our disposal, it has been assigned to be used for the mirrors and other volunteer systems. Each worker node mounts the same SAN over a high bandwidth backplane, so packages are consistent across them regardless of what worker node a user is sent to to serve the actual request.

Additional info:
Old DNS entries being replaced: mirror{1,2,3}.hackingand.coffee
New DNS Entries (either one works, they go to the same balancer): mirror[s].hackingand.coffee
Upstream has also been switched away from RIT.edu to Kernel.org, as we have recently been getting older packages in the mix or unpredictable rsync failures.
Old DNS entries will stay functional for the time being, however, they are going to point towards the new balancer instead which has a valid cert for each of them.

Cheers,
-Jim