I can't remember if I've asked this already, or if someone else has brought this up.

I have noticed that gluster replication in a hyperconverged environment is very slow.

I just (successfully, this time) added a brick to volume that was originally built on a single-node Hyperconverged cluster. 

One of my volumes has over 600GB, but the network is 10Gbps, so replication really shouldn't take that long.

It's been over two hours, and that volume has only replicated a little over 400GB. It feels like it's taking about 1GB every 30 seconds - 1 minute or so...

Why would this replication be so slow?