We have two BorderManager 3.8 servers across a S2S VPN.

At present we have some strange random NMAS errors occuring
when using C2S VPN to the server holding the Master Replica.

DSRepair shows zero errors on both servers, Timesync is also

I wondered if the NDS design was the problem. For
example we haven't created a top level Organisation (O) with
two Organisationl Units(OU) underneath. Instead we have
used two O's from the [Root]. The reason for doing this is Site A
and Site B are two different companies. I have also created a partition
on each O boundry.

As follows :

O=Site A (Master Replica) - Partition created
O=Site B (Read/Write Replica) - Partition created

Each site has a single BM 3.8 server.

We have just patched both servers to NW65SP3 and BM38SP3 but the
NMAS errors are still occuring. Typically, 0xFFFFF99D, 0xFFFFF98C and

There doesn't seem to be a pattern to the errors..

Any ideas?