Related to #3597

While working on reencryption #3651, we have determined that some of our calls may start to take a much longer time than previous operations we've added.

For example, in the reencryption case, operations may take a significant amount of time. This requires a write lock on the individual pool, but this results in a write_all lock blocking until the reencryption operation is complete. The end result of this is that pools that are not actively being operated on cannot be managed in the background until the reencryption operation completes. I think the best solution to this would be to add a write_all_free option to operate on all pools that are not currently locked for other purposes. This would allow us to manage the other pools in the background without blocking on reencryption operations and improve paralllelism.