Wait until DNS returns the expected number of IPs

cmd/root.go

@@ -4,8 +4,10 @@ import (
 	"context"
 	"fmt"
 	"math"
+	"net"
 	"os"
 	"os/signal"
+	"sort"
 	"strconv"
 	"strings"
 	"sync"
@@ -356,14 +358,36 @@ func join_cluster(expectedInstance int, serviceName string) {
 		os.Exit(1)
 	}
 
+	// wait until DNS name serviceName returns the expected number of instances
+	var nodes []net.IP
+	var err error
+	for {
+		nodes, err = net.LookupIP(serviceName)
+		if err != nil {
+			log.Error("failed to lookup DNS name; retrying...", "name", serviceName, "error", err)
+			time.Sleep(time.Second)
+			continue
+		}
+		if len(nodes) >= expectedInstance {
+			log.Info("reached the expected number of instances", "expected instances", expectedInstance, "current instances", len(nodes))
+			break
+		}
+		log.Info("waiting for DNS to return the expected number of IPs", "exepcted", expectedInstance, "current", len(nodes))
+		time.Sleep(time.Second)
+	}
+
 	list, err := memberlist.Create(memberlist.DefaultLANConfig())
 	if err != nil {
 		panic("Failed to create memberlist: " + err.Error())
 	}
 
+	sort.Slice(nodes, func(i, j int) bool {
+		return nodes[i].String() < nodes[j].String()
+	})
+
 	// join the cluster
 	for {
-		_, err = list.Join([]string{serviceName})
+		_, err = list.Join([]string{nodes[0].String()})
 		if err == nil {
 			break
 		}