Propagate message to a node only if the cluster link is healthy. (#11752)

Currently while a sharded pubsub message publish tries to propagate the message across the cluster, a NULL check is missing for clusterLink. clusterLink could be NULL if the link is causing memory beyond the set threshold cluster-link-sendbuf-limit and server terminates the link.

This change introduces two things:

Avoids the engine crashes on the publishing node if a message is tried to be sent to a node and the link is NULL.
Adds a debugging tool CLUSTERLINK KILL to terminate the clusterLink between two nodes.

src/cluster.c

@@ -3198,6 +3198,9 @@ void clusterReadHandler(connection *conn) {
  * the link to be invalidated, so it is safe to call this function
  * from event handlers that will do stuff with the same link later. */
 void clusterSendMessage(clusterLink *link, clusterMsgSendBlock *msgblock) {
+    if (!link) {
+        return;
+    }
     if (listLength(link->send_msg_queue) == 0 && msgblock->msg.totlen != 0)
         connSetWriteHandlerWithBarrier(link->conn, clusterWriteHandler, 1);
 
@@ -3228,7 +3231,6 @@ void clusterBroadcastMessage(clusterMsgSendBlock *msgblock) {
     while((de = dictNext(di)) != NULL) {
         clusterNode *node = dictGetVal(de);
 
-        if (!node->link) continue;
         if (node->flags & (CLUSTER_NODE_MYSELF|CLUSTER_NODE_HANDSHAKE))
             continue;
         clusterSendMessage(node->link,msgblock);
@@ -3622,9 +3624,9 @@ void clusterPropagatePublish(robj *channel, robj *message, int sharded) {
     msgblock = clusterCreatePublishMsgBlock(channel, message, CLUSTERMSG_TYPE_PUBLISHSHARD);
     while((ln = listNext(&li))) {
         clusterNode *node = listNodeValue(ln);
-        if (node != myself) {
-            clusterSendMessage(node->link,msgblock);
-        }
+        if (node->flags & (CLUSTER_NODE_MYSELF|CLUSTER_NODE_HANDSHAKE))
+            continue;
+        clusterSendMessage(node->link,msgblock);
     }
     clusterMsgSendBlockDecrRefCount(msgblock);
 }

src/cluster.h

@@ -424,5 +424,6 @@ void clusterUpdateMyselfHostname(void);
 void clusterUpdateMyselfAnnouncedPorts(void);
 sds clusterGenNodesDescription(int filter, int use_pport);
 sds genClusterInfoString();
+void freeClusterLink(clusterLink *link);
 
 #endif /* __CLUSTER_H */

src/debug.c

@@ -492,6 +492,8 @@ void debugCommand(client *c) {
 "    In case RESET is provided the peak reset time will be restored to the default value",
 "REPLYBUFFER RESIZING <0|1>",
 "    Enable or disable the reply buffer resize cron job",
+"CLUSTERLINK KILL <to|from|all> <node-id>",
+"    Kills the link based on the direction to/from (both) with the provided node." ,
 NULL
         };
         addReplyHelp(c, help);
@@ -997,6 +999,33 @@ NULL
             return;
         }
         addReply(c, shared.ok);
+    } else if(!strcasecmp(c->argv[1]->ptr,"CLUSTERLINK") &&
+        !strcasecmp(c->argv[2]->ptr,"KILL") &&
+        c->argc == 5) {
+        if (!server.cluster_enabled) {
+            addReplyError(c, "Debug option only available for cluster mode enabled setup!");
+            return;
+        }
+
+        /* Find the node. */
+        clusterNode *n = clusterLookupNode(c->argv[4]->ptr, sdslen(c->argv[4]->ptr));
+        if (!n) {
+            addReplyErrorFormat(c,"Unknown node %s", (char*)c->argv[4]->ptr);
+            return;
+        }
+
+        /* Terminate the link based on the direction or all. */
+        if (!strcasecmp(c->argv[3]->ptr,"from")) {
+            freeClusterLink(n->inbound_link);
+        } else if (!strcasecmp(c->argv[3]->ptr,"to")) {
+            freeClusterLink(n->link);
+        } else if (!strcasecmp(c->argv[3]->ptr,"all")) {
+            freeClusterLink(n->link);
+            freeClusterLink(n->inbound_link);
+        } else {
+            addReplyErrorFormat(c, "Unknown direction %s", (char*) c->argv[3]->ptr);
+        }
+        addReply(c,shared.ok);
     } else {
         addReplySubcommandSyntaxError(c);
         return;

tests/unit/cluster/links.tcl

@@ -67,6 +67,61 @@ proc publish_messages {server num_msgs msg_size} {
     }
 }
 
+start_cluster 1 2 {tags {external:skip cluster}} {
+    set primary_id 0
+    set replica1_id 1
+
+    set primary [Rn $primary_id]
+    set replica1 [Rn $replica1_id]
+
+    test "Broadcast message across a cluster shard while a cluster link is down" {
+        set replica1_node_id [$replica1 CLUSTER MYID]
+
+        set channelname ch3
+
+        # subscribe on replica1
+        set subscribeclient1 [redis_deferring_client -1]
+        $subscribeclient1 deferred 1
+        $subscribeclient1 SSUBSCRIBE $channelname
+        $subscribeclient1 read
+
+        # subscribe on replica2
+        set subscribeclient2 [redis_deferring_client -2]
+        $subscribeclient2 deferred 1
+        $subscribeclient2 SSUBSCRIBE $channelname
+        $subscribeclient2 read
+
+        # Verify number of links with cluster stable state
+        assert_equal [expr [number_of_peers $primary_id]*2] [number_of_links $primary_id]
+
+        # Disconnect the cluster between primary and replica1 and publish a message.
+        $primary MULTI
+        $primary DEBUG CLUSTERLINK KILL TO $replica1_node_id
+        $primary SPUBLISH $channelname hello
+        set res [$primary EXEC]
+
+        # Verify no client exists on the primary to receive the published message.
+        assert_equal $res {OK 0}
+
+        # Wait for all the cluster links are healthy
+        wait_for_condition 50 100 {
+            [number_of_peers $primary_id]*2 == [number_of_links $primary_id]
+        } else {
+            fail "All peer links couldn't be established"
+        }
+
+        # Publish a message afterwards.
+        $primary SPUBLISH $channelname world
+
+        # Verify replica1 has received only (world) / hello is lost.
+        assert_equal "smessage ch3 world" [$subscribeclient1 read]
+
+        # Verify replica2 has received both messages (hello/world)
+        assert_equal "smessage ch3 hello" [$subscribeclient2 read]
+        assert_equal "smessage ch3 world" [$subscribeclient2 read]
+    } {} {needs:debug}
+}
+
 start_cluster 3 0 {tags {external:skip cluster}} {
     test "Each node has two links with each peer" {
         for {set id 0} {$id < [llength $::servers]} {incr id} {
@@ -161,7 +216,6 @@ start_cluster 3 0 {tags {external:skip cluster}} {
         } else {
             fail "Cluster link not freed as expected"
         }
-        puts -nonewline "$i 128KB messages needed to overflow 256KB buffer limit. "
 
         # A new link to primary2 should have been recreated
         set new_link_p1_to_p2 [get_link_to_peer $primary1_id $primary2_name]