From e13b671aa510f553f4a6a232b4694b6a4cce88ae Mon Sep 17 00:00:00 2001
From: Jason Lowe <jlowe@apache.org>
Date: Wed, 3 Jun 2015 19:44:07 +0000
Subject: [PATCH] YARN-3585. NodeManager cannot exit on SHUTDOWN event
 triggered and NM recovery is enabled. Contributed by Rohith Sharmaks

---
 hadoop-yarn-project/CHANGES.txt                 |  3 +++
 .../yarn/server/nodemanager/NodeManager.java    | 17 +++++++++++++++--
 2 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/hadoop-yarn-project/CHANGES.txt b/hadoop-yarn-project/CHANGES.txt
index c5ba58c3cf5..64f6abbd07a 100644
--- a/hadoop-yarn-project/CHANGES.txt
+++ b/hadoop-yarn-project/CHANGES.txt
@@ -615,6 +615,9 @@ Release 2.7.1 - UNRELEASED
     YARN-3725. App submission via REST API is broken in secure mode due to 
     Timeline DT service address is empty. (Zhijie Shen via wangda)
 
+    YARN-3585. NodeManager cannot exit on SHUTDOWN event triggered and NM
+    recovery is enabled (Rohith Sharmaks via jlowe)
+
 Release 2.7.0 - 2015-04-20
 
   INCOMPATIBLE CHANGES
diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeManager.java b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeManager.java
index 03e17c83a79..185ba126a4d 100644
--- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeManager.java
+++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeManager.java
@@ -38,6 +38,7 @@ import org.apache.hadoop.metrics2.lib.DefaultMetricsSystem;
 import org.apache.hadoop.security.Credentials;
 import org.apache.hadoop.security.SecurityUtil;
 import org.apache.hadoop.service.CompositeService;
+import org.apache.hadoop.util.ExitUtil;
 import org.apache.hadoop.util.GenericOptionsParser;
 import org.apache.hadoop.util.NodeHealthScriptRunner;
 import org.apache.hadoop.util.ReflectionUtils;
@@ -94,6 +95,7 @@ public class NodeManager extends CompositeService
   
   private AtomicBoolean isStopping = new AtomicBoolean(false);
   private boolean rmWorkPreservingRestartEnabled;
+  private boolean shouldExitOnShutdownEvent = false;
 
   public NodeManager() {
     super(NodeManager.class.getName());
@@ -344,7 +346,16 @@ public class NodeManager extends CompositeService
     new Thread() {
       @Override
       public void run() {
-        NodeManager.this.stop();
+        try {
+          NodeManager.this.stop();
+        } catch (Throwable t) {
+          LOG.error("Error while shutting down NodeManager", t);
+        } finally {
+          if (shouldExitOnShutdownEvent
+              && !ShutdownHookManager.get().isShutdownInProgress()) {
+            ExitUtil.terminate(-1);
+          }
+        }
       }
     }.start();
   }
@@ -530,7 +541,9 @@ public class NodeManager extends CompositeService
       nodeManagerShutdownHook = new CompositeServiceShutdownHook(this);
       ShutdownHookManager.get().addShutdownHook(nodeManagerShutdownHook,
                                                 SHUTDOWN_HOOK_PRIORITY);
-
+      // System exit should be called only when NodeManager is instantiated from
+      // main() funtion
+      this.shouldExitOnShutdownEvent = true;
       this.init(conf);
       this.start();
     } catch (Throwable t) {