HCOLL-134 changed the default so that is does not autoReconnectedUponDroppedConnection, however it better if most of the test cases have this can you can guarantee that the server is ready to accept TCP/IP before the client.

src/main/java/net/openhft/chronicle/map/NodeDiscovery.java

@@ -172,7 +172,7 @@ public class NodeDiscovery {
         // the identifiers will come back to the callback on the nio thread, the update arrives at the
         // onRemoteNodeEvent
 
-        Thread.sleep(200);
+        Thread.sleep(1000);
 
         byte identifier;
 
@@ -191,7 +191,7 @@ public class NodeDiscovery {
             final ProposedNodes proposedNodes = new
                     ProposedNodes(ourAddressAndPort, identifier);
 
-            Thread.sleep(100);
+            Thread.sleep(1000);
 
             countDownLatch.set(new CountDownLatch(1));
 
@@ -245,16 +245,18 @@ public class NodeDiscovery {
         // add our identifier and host:port to the list of known identifiers
         knownNodes.add(ourAddressAndPort, identifier);
 
-        final TcpTransportAndNetworkConfig tcpConfig = TcpTransportAndNetworkConfig
+    /*  final TcpTransportAndNetworkConfig tcpConfig = TcpTransportAndNetworkConfig
                 .of(ourAddressAndPort.getPort(),
-                        toInetSocketCollection(knownHostPorts0))
-
-                .heartBeatInterval(1, SECONDS);
-
+                        toInetSocketCollection(knownHostPorts0));
+*/
 
         LOG.info("Using Remote identifier=" + identifier);
         nodeDiscoveryEventListenerAtomicReference.set(null);
 
+
+        TcpTransportAndNetworkConfig tcpConfig = TcpTransportAndNetworkConfig.of(8123, new
+                InetSocketAddress("192.168.1.253", 8123));
+
         final ReplicationHub replicationHub = ReplicationHub.builder()
                 .maxEntrySize(10 * 1024)
                 .tcpTransportAndNetwork(tcpConfig)