We have cleaned up our storage (CS-1431). We wrote these three steps:

1) We clean up c8s and c9s content. Using both pruning and garbage collection. At least to a point it'll work for a few months. Doesn't need to be automated or final, we just need time. (done)

2) We put servers using the mount on ansible. Mount it everywhere. And configure Koji to use the new volume for c10s. This unblocks c10s work. Which is why we do 1) quickly as a one off, so we don't wait with c10s for too long. (CS-1510)

3) We develop proper garbage collection and pruning policy, so we're good long-term. (this ticket)

Basically, we did a one time cleanup, now we need to automate it.