Cluster: Assign default human-readable node name when starting server

src/cluster_legacy.c

@@ -1215,7 +1215,11 @@ static void updateAnnouncedHostname(clusterNode *node, char *value) {
 }
 
 static void updateAnnouncedHumanNodename(clusterNode *node, char *value) {
-    updateSdsExtensionField(&node->human_nodename, value);
+    /* We should only update the human nodename when the provided new
+     * value isn't NULL, otherwise the following function will clear
+     * the human nodename field. */
+    if (value != NULL)
+        updateSdsExtensionField(&node->human_nodename, value);
 }
 
 static void updateAnnouncedClientIpV4(clusterNode *node, char *value) {
@@ -1284,6 +1288,13 @@ static void updateShardId(clusterNode *node, const char *shard_id) {
     }
 }
 
+static void updateHumanNodenameToAddress(clusterNode *node) {
+    const int port = server.tls_cluster ? node->tls_port : node->tcp_port;
+    char buf[64];
+    snprintf(buf, sizeof(buf), "%s:%d", node->ip, port);
+    updateAnnouncedHumanNodename(node, buf);
+}
+
 static inline int areInSameShard(clusterNode *node1, clusterNode *node2) {
     return memcmp(node1->shard_id, node2->shard_id, CLUSTER_NAMELEN) == 0;
 }
@@ -1408,7 +1419,18 @@ void clusterInit(void) {
     clusterUpdateMyselfClientIpV4();
     clusterUpdateMyselfClientIpV6();
     clusterUpdateMyselfHostname();
-    clusterUpdateMyselfHumanNodename();
+    /* Assigning human-readable nodename at start-up makes it much
+     * easier to read the server logs. By default, a server doesn't
+     * have a human-readable nodename unless explicitly assigned by
+     * CONFIG SET command or config file edit, so we simply use
+     * the server's IP and port as the nodename. This name will be
+     * carried in the PING extension so that all nodes in the cluster
+     * will know this name eventually. */
+    if (server.cluster_announce_human_nodename != NULL &&
+        server.cluster_announce_human_nodename[0] != '\0')
+        clusterUpdateMyselfHumanNodename();
+    else
+        updateHumanNodenameToAddress(myself);
     resetClusterStats();
 }
 
@@ -3746,6 +3768,7 @@ int clusterProcessPacket(clusterLink *link) {
                 memcpy(myself->ip, ip, NET_IP_STR_LEN);
                 serverLog(LL_NOTICE, "IP address for this node updated to %s", myself->ip);
                 clusterDoBeforeSleep(CLUSTER_TODO_SAVE_CONFIG);
+                updateHumanNodenameToAddress(myself);
             }
         }
 
@@ -5823,7 +5846,7 @@ void clusterCron(void) {
     long long cluster_node_conn_attempts = maxConnectionAttemptsPerCron();
     while ((de = dictNext(di)) != NULL) {
         clusterNode *node = dictGetVal(de);
-        /* We free the inbound or outboud link to the node if the link has an
+        /* We free the inbound or outbound link to the node if the link has an
          * oversized message send queue and immediately try reconnecting. */
         clusterNodeCronFreeLinkOnBufferLimitReached(node);
         /* The protocol is that function(s) below return non-zero if the node was

tests/support/util.tcl

@@ -201,7 +201,7 @@ proc verify_log_message {srv_idx pattern from_line} {
     incr from_line
     set result [exec tail -n +$from_line < [srv $srv_idx stdout]]
     if {![string match $pattern $result]} {
-        fail "expected message not found in log file: $pattern"
+        fail "expected pattern found in log file: $pattern, but instead got: $result"
     }
 }
 

tests/unit/cluster/manual-failover.tcl

@@ -418,6 +418,9 @@ start_cluster 3 1 {tags {external:skip cluster}} {
         set R2_nodeid [R 2 cluster myid]
         set R3_nodeid [R 3 cluster myid]
 
+        set R0_port [srv 0 port]
+        set R3_port [srv -3 port]
+
         set R0_shardid [R 0 cluster myshardid]
         set R3_shardid [R 3 cluster myshardid]
         assert_equal $R0_shardid $R3_shardid
@@ -478,21 +481,21 @@ start_cluster 3 1 {tags {external:skip cluster}} {
         # verify we print the logs.
 
         # Both importing slots and migrating slots are move to R3.
-        set pattern "*Failover occurred in migration source. Update importing source for slot 0 to node $R3_nodeid () in shard $R3_shardid*"
+        set pattern "*Failover occurred in migration source. Update importing source for slot 0 to node $R3_nodeid (127.0.0.1:$R3_port) in shard $R3_shardid*"
         verify_log_message -1 $pattern $loglines1
-        set pattern "*Failover occurred in migration target. Slot 5462 is now being migrated to node $R3_nodeid () in shard $R3_shardid*"
+        set pattern "*Failover occurred in migration target. Slot 5462 is now being migrated to node $R3_nodeid (127.0.0.1:$R3_port) in shard $R3_shardid*"
         verify_log_message -1 $pattern $loglines1
 
         # Both slots are move to R3.
         set R0_slots 5462
-        set pattern "*A failover occurred in shard $R3_shardid; node $R0_nodeid () lost $R0_slots slot(s) and failed over to node $R3_nodeid*"
+        set pattern "*A failover occurred in shard $R3_shardid; node $R0_nodeid (127.0.0.1:$R0_port) lost $R0_slots slot(s) and failed over to node $R3_nodeid*"
         verify_log_message -1 $pattern $loglines1
         verify_log_message -2 $pattern $loglines2
 
         # Both importing slots and migrating slots are move to R3.
-        set pattern "*A failover occurred in migration source. Update importing source of 1 slot(s) to node $R3_nodeid () in shard $R3_shardid*"
+        set pattern "*A failover occurred in migration source. Update importing source of 1 slot(s) to node $R3_nodeid (127.0.0.1:$R3_port) in shard $R3_shardid*"
         verify_log_message -1 $pattern $loglines1
-        set pattern "*A failover occurred in migration target. Update migrating target of 1 slot(s) to node $R3_nodeid () in shard $R3_shardid*"
+        set pattern "*A failover occurred in migration target. Update migrating target of 1 slot(s) to node $R3_nodeid (127.0.0.1:$R3_port) in shard $R3_shardid*"
         verify_log_message -1 $pattern $loglines1
 
         R 1 debug disable-cluster-reconnection 0