HDFS-6089. Standby NN while transitioning to active throws a connection refused error when the prior active NN process is suspended. Contributed by Jing Zhao.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-2@1579694 13f79535-47bb-0310-9956-ffa450edef68

hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt

@@ -482,6 +482,10 @@ Release 2.4.0 - UNRELEASED
     HDFS-6107. When a block can't be cached due to limited space on the
     DataNode, that block becomes uncacheable (cmccabe)
 
+    HDFS-6089. Standby NN while transitioning to active throws a connection
+    refused error when the prior active NN process is suspended.
+    (Jing Zhao via Andrew Wang)
+
   BREAKDOWN OF HDFS-4685 SUBTASKS AND RELATED JIRAS
 
     HDFS-5596. Implement RPC stubs. (Haohui Mai via cnauroth)

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java

@@ -524,6 +524,8 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
   public static final int DFS_HA_LOGROLL_PERIOD_DEFAULT = 2 * 60; // 2m
   public static final String DFS_HA_TAILEDITS_PERIOD_KEY = "dfs.ha.tail-edits.period";
   public static final int DFS_HA_TAILEDITS_PERIOD_DEFAULT = 60; // 1m
+  public static final String DFS_HA_LOGROLL_RPC_TIMEOUT_KEY = "dfs.ha.log-roll.rpc.timeout";
+  public static final int DFS_HA_LOGROLL_RPC_TIMEOUT_DEFAULT = 20000; // 20s
   public static final String DFS_HA_FENCE_METHODS_KEY = "dfs.ha.fencing.methods";
   public static final String DFS_HA_AUTO_FAILOVER_ENABLED_KEY = "dfs.ha.automatic-failover.enabled";
   public static final boolean DFS_HA_AUTO_FAILOVER_ENABLED_DEFAULT = false;

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/ha/EditLogTailer.java

@@ -135,9 +135,12 @@ public class EditLogTailer {
   
   private NamenodeProtocol getActiveNodeProxy() throws IOException {
     if (cachedActiveProxy == null) {
-      NamenodeProtocolPB proxy = 
-        RPC.waitForProxy(NamenodeProtocolPB.class,
-            RPC.getProtocolVersion(NamenodeProtocolPB.class), activeAddr, conf);
+      int rpcTimeout = conf.getInt(
+          DFSConfigKeys.DFS_HA_LOGROLL_RPC_TIMEOUT_KEY,
+          DFSConfigKeys.DFS_HA_LOGROLL_RPC_TIMEOUT_DEFAULT);
+      NamenodeProtocolPB proxy = RPC.waitForProxy(NamenodeProtocolPB.class,
+          RPC.getProtocolVersion(NamenodeProtocolPB.class), activeAddr, conf,
+          rpcTimeout, Long.MAX_VALUE);
       cachedActiveProxy = new NamenodeProtocolTranslatorPB(proxy);
     }
     assert cachedActiveProxy != null;