This alarm occurs when DSA A tries to synchronize DSA B using MW-DISP after start-up/outage, and the connection attempt (DSP BIND) is refused. Recovery (MW-DISP) is then attempted every 1 minute to try and make a successful connection to DSA B. The error however is only displayed every 10th iteration to save a bit of junking up the alarm log during an extended outage.
To determine the cause, you will need to check the connectivity between DSA A and DSA B to see why the bind is being refused. Maybe the alarm/warn log of DSA B has a clue? This kind of issue is typically configuration related.
I'm surprised that replication is working. Replication doesn't start until recovery has been successful, otherwise, the DSA will be replicating over the top of stale data. If you telnet to DXconsole and run "get dsp;" this will show the status of replication. A status of "OK" on both DSA A and DSA B indicate the recovery has completed successfully and replication is working as expected.