1 master- edir server and 4 servers where the edir is replicated to, from which 2 are cluster nodes.

Master-edir is the only Directory Agent (SLP).

If we restart the master, the clients cannot mount their drives and the volume on the cluster goes comatose.
We had to restart the nodes and wait a while

We think that this is because the master is the only DA and if you restart it in the running system it leads to that problem.

1. Is this correct?
2. Can /should we make another server a DA so we won't have this "single point of failure"?