Fix a potential risk to send multiple http requests to global cluster endpoint in short time

Signed-off-by: yhmo <yihua.mo@zilliz.com>

Author: yhmo

sdk-core/src/main/java/io/milvus/v2/client/globalcluster/TopologyRefresher.java

@@ -25,6 +25,7 @@
 import java.util.concurrent.Executors;
 import java.util.concurrent.ScheduledExecutorService;
 import java.util.concurrent.TimeUnit;
+import java.util.concurrent.atomic.AtomicBoolean;
 import java.util.concurrent.atomic.AtomicLong;
 import java.util.function.Consumer;
 
@@ -37,6 +38,7 @@ public class TopologyRefresher {
     private final Consumer<GlobalTopology> onTopologyChange;
     private final ScheduledExecutorService scheduler;
     private final AtomicLong currentVersion;
+    private final AtomicBoolean refreshing = new AtomicBoolean(false);
 
     public TopologyRefresher(String globalEndpoint, String token, long initialVersion,
                              Consumer<GlobalTopology> onTopologyChange) {
@@ -59,14 +61,26 @@ public void start() {
     }
 
     public void triggerRefresh() {
-        scheduler.submit(this::refresh);
+        if (refreshing.getAndSet(true)) {
+            logger.debug("Topology refresh already in progress, skipping");
+            return;
+        }
+        scheduler.submit(this::refreshWithCleanup);
     }
 
     public void stop() {
         scheduler.shutdownNow();
         logger.info("Global topology refresher stopped for endpoint: {}", globalEndpoint);
     }
 
+    private void refreshWithCleanup() {
+        try {
+            refresh();
+        } finally {
+            refreshing.set(false);
+        }
+    }
+
     private void refresh() {
         try {
             GlobalTopology topology = GlobalClusterUtils.fetchTopology(globalEndpoint, token);