Fly volumes are neither distributed nor replicated. Fly haven’t commit to building one, either.
Not really the kind of replication you’re looking for, but you can consider running Ceph with BlueStore and cluster it up as a Block Device with read-cache. I am not aware of anyone running Ceph on Fly. And it frankly sounds non-trivial to do so.
Btw, here’s an example Fly app running a MinIO cluster (a S3 clone). Then, there’s this poor-man’s Fly-clone, Apache Helix, which bundles in a replicated fs, but seems like an overkill to run Helix on Fly!
You could also get fancy with other rsync-based replication setups (like volsync / orifs / sfs); tbh, I’d be wary of using those in production.
Someone’s running IPFS on Fly (which functions like git+rsync), so that’s there too.