It looks like that might have helped a little. One of the three servers are converting now, Server 2 with 8x2TB disks. The other two fails like before.
I've attached the worker log from one of the failed servers (Server1) and from the one that is working (Server2).