HDDS-3498. Shutdown datanode if address is already in use (#7256)

apache · Oct 4, 2024 · 30da31f · 30da31f
1 parent 2401d27
commit 30da31f
Show file tree

Hide file tree

Showing 3 changed files with 23 additions and 7 deletions.
diff --git a/...main/java/org/apache/hadoop/ozone/container/common/statemachine/EndpointStateMachine.java b/...main/java/org/apache/hadoop/ozone/container/common/statemachine/EndpointStateMachine.java
@@ -234,12 +234,17 @@ public void logIfNeeded(Exception ex) {
     }
 
     if (missCounter == 0) {
+      long missedDurationSeconds = TimeUnit.MILLISECONDS.toSeconds(
+              this.getMissedCount() * getScmHeartbeatInterval(this.conf)
+      );
       LOG.warn(
-          "Unable to communicate to {} server at {} for past {} seconds.",
-          serverName,
-          getAddress().getHostString() + ":" + getAddress().getPort(),
-          TimeUnit.MILLISECONDS.toSeconds(this.getMissedCount() *
-                  getScmHeartbeatInterval(this.conf)), ex);
+              "Unable to communicate to {} server at {}:{} for past {} seconds.",
+              serverName,
+              address.getAddress(),
+              address.getPort(),
+              missedDurationSeconds,
+              ex
+      );
     }
 
     if (LOG.isTraceEnabled()) {

diff --git a/...in/java/org/apache/hadoop/ozone/container/common/states/endpoint/VersionEndpointTask.java b/...in/java/org/apache/hadoop/ozone/container/common/states/endpoint/VersionEndpointTask.java
@@ -17,6 +17,7 @@
 package org.apache.hadoop.ozone.container.common.states.endpoint;
 
 import java.io.IOException;
+import java.net.BindException;
 import java.util.concurrent.Callable;
 
 import org.apache.hadoop.hdds.conf.ConfigurationSource;
@@ -104,7 +105,7 @@ public EndpointStateMachine.EndPointStates call() throws Exception {
         LOG.debug("Cannot execute GetVersion task as endpoint state machine " +
             "is in {} state", rpcEndPoint.getState());
       }
-    } catch (DiskOutOfSpaceException ex) {
+    } catch (DiskOutOfSpaceException | BindException ex) {
       rpcEndPoint.setState(EndpointStateMachine.EndPointStates.SHUTDOWN);
     } catch (IOException ex) {
       rpcEndPoint.logIfNeeded(ex);

diff --git a/...ain/java/org/apache/hadoop/ozone/container/common/transport/server/XceiverServerGrpc.java b/...ain/java/org/apache/hadoop/ozone/container/common/transport/server/XceiverServerGrpc.java
@@ -19,6 +19,7 @@
 package org.apache.hadoop.ozone.container.common.transport.server;
 
 import java.io.IOException;
+import java.net.BindException;
 import java.util.Collections;
 import java.util.List;
 import java.util.UUID;
@@ -185,7 +186,16 @@ public HddsProtos.ReplicationType getServerType() {
   @Override
   public void start() throws IOException {
     if (!isStarted) {
-      server.start();
+      try {
+        server.start();
+      } catch (IOException e) {
+        LOG.error("Error while starting the server", e);
+        if (e.getMessage().contains("Failed to bind to address")) {
+          throw new BindException(e.getMessage());
+        } else {
+          throw e;
+        }
+      }
       int realPort = server.getPort();
 
       if (port == 0) {