svn merge -c 1354798 from branch-1 for HDFS-3518.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-1.1@1354799 13f79535-47bb-0310-9956-ffa450edef68

CHANGES.txt

@@ -25,6 +25,9 @@ Release 1.1.0 - unreleased
 
     HADOOP-8209. Add option to relax build-version check for branch-1. (eli)
 
+    HDFS-3518. Add a utility method DistributedFileSystem.isHealthy(uri) for
+    checking if the given HDFS is healthy. (szetszwo)
+
   IMPROVEMENTS
 
     MAPREDUCE-3597. [Rumen] Provide a way to access other info of history file

src/hdfs/org/apache/hadoop/hdfs/DistributedFileSystem.java

@@ -18,29 +18,40 @@
 
 package org.apache.hadoop.hdfs;
 
-import java.io.*;
-import java.net.*;
+import java.io.FileNotFoundException;
+import java.io.IOException;
+import java.net.InetSocketAddress;
+import java.net.URI;
 import java.util.ArrayList;
 
-import org.apache.hadoop.fs.permission.FsPermission;
-import org.apache.hadoop.fs.*;
 import org.apache.hadoop.conf.Configuration;
-import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
-import org.apache.hadoop.hdfs.protocol.FSConstants;
+import org.apache.hadoop.fs.BlockLocation;
+import org.apache.hadoop.fs.ContentSummary;
+import org.apache.hadoop.fs.FSDataInputStream;
+import org.apache.hadoop.fs.FSDataOutputStream;
+import org.apache.hadoop.fs.FileStatus;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.MD5MD5CRC32FileChecksum;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.fs.permission.FsPermission;
 import org.apache.hadoop.hdfs.protocol.Block;
-import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
-import org.apache.hadoop.hdfs.protocol.LocatedBlock;
+import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
 import org.apache.hadoop.hdfs.protocol.DirectoryListing;
+import org.apache.hadoop.hdfs.protocol.FSConstants;
 import org.apache.hadoop.hdfs.protocol.FSConstants.DatanodeReportType;
+import org.apache.hadoop.hdfs.protocol.FSConstants.SafeModeAction;
 import org.apache.hadoop.hdfs.protocol.FSConstants.UpgradeAction;
+import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
+import org.apache.hadoop.hdfs.protocol.LocatedBlock;
 import org.apache.hadoop.hdfs.security.token.delegation.DelegationTokenIdentifier;
 import org.apache.hadoop.hdfs.server.common.UpgradeStatusReport;
 import org.apache.hadoop.hdfs.server.namenode.NameNode;
-import org.apache.hadoop.hdfs.DFSClient.DFSOutputStream;
+import org.apache.hadoop.io.IOUtils;
 import org.apache.hadoop.io.Text;
+import org.apache.hadoop.ipc.Client;
 import org.apache.hadoop.security.AccessControlException;
-import org.apache.hadoop.security.token.Token;
 import org.apache.hadoop.security.token.SecretManager.InvalidToken;
+import org.apache.hadoop.security.token.Token;
 import org.apache.hadoop.util.Progressable;
 
 
@@ -626,4 +637,46 @@ public class DistributedFileSystem extends FileSystem {
   public void setBalancerBandwidth(long bandwidth) throws IOException {
     dfs.setBalancerBandwidth(bandwidth);
   }
+
+  /**
+   * Is the HDFS healthy?
+   * HDFS is considered as healthy if it is up and not in safemode.
+   *
+   * @param uri the HDFS URI.  Note that the URI path is ignored.
+   * @return true if HDFS is healthy; false, otherwise.
+   */
+  public static boolean isHealthy(URI uri) {
+    //check scheme
+    final String scheme = uri.getScheme();
+    if (!"hdfs".equalsIgnoreCase(scheme)) {
+      throw new IllegalArgumentException("This scheme is not hdfs, uri=" + uri);
+    }
+    
+    final Configuration conf = new Configuration();
+    //disable FileSystem cache
+    conf.setBoolean(String.format("fs.%s.impl.disable.cache", scheme), true);
+    //disable client retry for rpc connection and rpc calls
+    conf.setBoolean(DFSConfigKeys.DFS_CLIENT_RETRY_POLICY_ENABLED_KEY, false);
+    conf.setInt(Client.IPC_CLIENT_CONNECT_MAX_RETRIES_KEY, 0);
+
+    DistributedFileSystem fs = null;
+    try {
+      fs = (DistributedFileSystem)FileSystem.get(uri, conf);
+      final boolean safemode = fs.setSafeMode(SafeModeAction.SAFEMODE_GET);
+      if (LOG.isDebugEnabled()) {
+        LOG.debug("Is namenode in safemode? " + safemode + "; uri=" + uri);
+      }
+
+      fs.close();
+      fs = null;
+      return !safemode;
+    } catch(IOException e) {
+      if (LOG.isDebugEnabled()) {
+        LOG.debug("Got an exception for uri=" + uri, e);
+      }
+      return false;
+    } finally {
+      IOUtils.cleanup(LOG, fs);
+    }
+  }
 }

src/test/org/apache/hadoop/hdfs/TestDFSClientRetries.java

@@ -28,6 +28,7 @@ import java.io.InputStream;
 import java.io.OutputStream;
 import java.net.InetSocketAddress;
 import java.net.SocketTimeoutException;
+import java.net.URI;
 import java.util.ArrayList;
 import java.util.List;
 import java.util.concurrent.TimeUnit;
@@ -542,9 +543,11 @@ public class TestDFSClientRetries extends TestCase {
         conf, numDatanodes, true, null);
     try {
       cluster.waitActive();
+      final DistributedFileSystem dfs = (DistributedFileSystem)cluster.getFileSystem();
+      final URI uri = dfs.getUri();
+      assertTrue(DistributedFileSystem.isHealthy(uri));
 
       //create a file
-      final DistributedFileSystem dfs = (DistributedFileSystem)cluster.getFileSystem();
       final long length = 1L << 20;
       final Path file1 = new Path(dir, "foo"); 
       DFSTestUtil.createFile(dfs, file1, length, numDatanodes, 20120406L);
@@ -554,7 +557,9 @@ public class TestDFSClientRetries extends TestCase {
       assertEquals(length, s1.getLen());
 
       //shutdown namenode
+      assertTrue(DistributedFileSystem.isHealthy(uri));
       cluster.shutdownNameNode();
+      assertFalse(DistributedFileSystem.isHealthy(uri));
 
       //namenode is down, create another file in a thread
       final Path file3 = new Path(dir, "file"); 
@@ -579,8 +584,10 @@ public class TestDFSClientRetries extends TestCase {
           try {
             //sleep, restart, and then wait active
             TimeUnit.SECONDS.sleep(30);
+            assertFalse(DistributedFileSystem.isHealthy(uri));
             cluster.restartNameNode(false, false);
             cluster.waitActive();
+            assertTrue(DistributedFileSystem.isHealthy(uri));
           } catch (Exception e) {
             exceptions.add(e);
           }
@@ -596,7 +603,9 @@ public class TestDFSClientRetries extends TestCase {
       assertEquals(dfs.getFileChecksum(file1), dfs.getFileChecksum(file3));
 
       //enter safe mode
+      assertTrue(DistributedFileSystem.isHealthy(uri));
       dfs.setSafeMode(SafeModeAction.SAFEMODE_ENTER);
+      assertFalse(DistributedFileSystem.isHealthy(uri));
       
       //leave safe mode in a new thread
       new Thread(new Runnable() {
@@ -605,7 +614,9 @@ public class TestDFSClientRetries extends TestCase {
           try {
             //sleep and then leave safe mode
             TimeUnit.SECONDS.sleep(30);
+            assertFalse(DistributedFileSystem.isHealthy(uri));
             dfs.setSafeMode(SafeModeAction.SAFEMODE_LEAVE);
+            assertTrue(DistributedFileSystem.isHealthy(uri));
           } catch (Exception e) {
             exceptions.add(e);
           }
@@ -616,6 +627,8 @@ public class TestDFSClientRetries extends TestCase {
       final Path file2 = new Path(dir, "bar");
       DFSTestUtil.createFile(dfs, file2, length, numDatanodes, 20120406L);
       assertEquals(dfs.getFileChecksum(file1), dfs.getFileChecksum(file2));
+
+      assertTrue(DistributedFileSystem.isHealthy(uri));
       
       //make sure it won't retry on exceptions like FileNotFoundException
       final Path nonExisting = new Path(dir, "nonExisting");