Replication issues after vcenter crash

[PaulNSW]

Our old v6.5 vcenter running on Windows, crapped out after the SQLExpress database hit its limit. One error lead to another, and it ended up being quicker to set up a new vcenter. This time we used the VM appliance version, reconnected the hosts, then used the migration script to update Veeam with the new vcenter details.

Backups started working ok, but replications kept failing. We ended up deleting all replicas from our remote site, then trying to replicate over again. For the smaller VMs, that went fine. The VMs over 5TB had a little trouble, "Digests finalisation has failed. Timeout reached," but after a few retries, they also replicated successfully.

However for the majority, incremental replications fail with "unable to parallel process" with parallel processing turned on, or else it cannot access the replica metadata because "the file is being used by another process." It's not being used by another process!

For the replication jobs that are working, the 1st disk replicates at full speed 44MB/s over 500mbit line, but subsequent disks get slower and slower, all the way down to 700KB/s

I've been going back and forth with support for the last month, and don't feel any closer to a solution
Case: 02111548

We've tried storing the metadata on different VMs and storage. Backup copy over the same link is a consistent speed. It just seems to be replications.

Are there any known issues with the latest 6.5b vcenter and replication? Everything else is the same, proxies, OS, storage, line speed - the hosts are still ESXI6.0 5050593 , proxies Server 2012 R2