Retry if node is closed; close nodes with delay

cluster.go

@@ -226,7 +226,7 @@ func (c *clusterNodes) NextGeneration() uint32 {
 }
 
 // GC removes unused nodes.
-func (c *clusterNodes) GC(generation uint32) error {
+func (c *clusterNodes) GC(generation uint32) {
 	var collected []*clusterNode
 	c.mu.Lock()
 	for i := 0; i < len(c.addrs); {
@@ -243,14 +243,11 @@ func (c *clusterNodes) GC(generation uint32) error {
 	}
 	c.mu.Unlock()
 
-	var firstErr error
+	time.AfterFunc(time.Minute, func() {
 		for _, node := range collected {
-		if err := node.Client.Close(); err != nil && firstErr == nil {
+			_ = node.Client.Close()
-			firstErr = err
 		}
-	}
+	})
-
-	return firstErr
 }
 
 func (c *clusterNodes) All() ([]*clusterNode, error) {
@@ -596,6 +593,10 @@ func (c *ClusterClient) Watch(fn func(*Tx) error, keys ...string) error {
 			break
 		}
 
+		if internal.IsRetryableError(err, true) {
+			continue
+		}
+
 		moved, ask, addr := internal.IsMovedError(err)
 		if moved || ask {
 			c.lazyReloadState()
@@ -606,6 +607,13 @@ func (c *ClusterClient) Watch(fn func(*Tx) error, keys ...string) error {
 			continue
 		}
 
+		if err == pool.ErrClosed {
+			node, err = state.slotMasterNode(slot)
+			if err != nil {
+				return err
+			}
+		}
+
 		return err
 	}
 
@@ -641,10 +649,10 @@ func (c *ClusterClient) Process(cmd Cmder) error {
 
 		if ask {
 			pipe := node.Client.Pipeline()
-			pipe.Process(NewCmd("ASKING"))
+			_ = pipe.Process(NewCmd("ASKING"))
-			pipe.Process(cmd)
+			_ = pipe.Process(cmd)
 			_, err = pipe.Exec()
-			pipe.Close()
+			_ = pipe.Close()
 			ask = false
 		} else {
 			err = node.Client.Process(cmd)
@@ -685,6 +693,14 @@ func (c *ClusterClient) Process(cmd Cmder) error {
 			continue
 		}
 
+		if err == pool.ErrClosed {
+			_, node, err = c.cmdSlotAndNode(state, cmd)
+			if err != nil {
+				cmd.setErr(err)
+				return err
+			}
+		}
+
 		break
 	}
 
@@ -921,7 +937,11 @@ func (c *ClusterClient) pipelineExec(cmds []Cmder) error {
 		for node, cmds := range cmdsMap {
 			cn, _, err := node.Client.getConn()
 			if err != nil {
+				if err == pool.ErrClosed {
+					c.remapCmds(cmds, failedCmds)
+				} else {
 					setCmdsErr(cmds, err)
+				}
 				continue
 			}
 
@@ -961,6 +981,18 @@ func (c *ClusterClient) mapCmdsByNode(cmds []Cmder) (map[*clusterNode][]Cmder, e
 	return cmdsMap, nil
 }
 
+func (c *ClusterClient) remapCmds(cmds []Cmder, failedCmds map[*clusterNode][]Cmder) {
+	remappedCmds, err := c.mapCmdsByNode(cmds)
+	if err != nil {
+		setCmdsErr(cmds, err)
+		return
+	}
+
+	for node, cmds := range remappedCmds {
+		failedCmds[node] = cmds
+	}
+}
+
 func (c *ClusterClient) pipelineProcessCmds(
 	node *clusterNode, cn *pool.Conn, cmds []Cmder, failedCmds map[*clusterNode][]Cmder,
 ) error {
@@ -1067,7 +1099,11 @@ func (c *ClusterClient) txPipelineExec(cmds []Cmder) error {
 			for node, cmds := range cmdsMap {
 				cn, _, err := node.Client.getConn()
 				if err != nil {
+					if err == pool.ErrClosed {
+						c.remapCmds(cmds, failedCmds)
+					} else {
 						setCmdsErr(cmds, err)
+					}
 					continue
 				}