Mailing-List: contact commits-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hbase.apache.org
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: svn commit: r1458531 - in /hbase/trunk/hbase-server/src:
 main/java/org/apache/hadoop/hbase/mapreduce/
 main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/
 test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/
Date: Tue, 19 Mar 2013 21:54:00 -0000
To: commits@hbase.apache.org
From: stack@apache.org
Message-Id: <20130319215400.C5D0623888CD@eris.apache.org>

Author: stack
Date: Tue Mar 19 21:54:00 2013
New Revision: 1458531

URL: http://svn.apache.org/r1458531
Log:
HBASE-8140 TableMapReduceUtils#addDependencyJar fails when nested inside another MR job

Added:
    hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/
    hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/JarFinder.java
    hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/
    hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/TestJarFinder.java
Modified:
    hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java

Modified: hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java
URL: http://svn.apache.org/viewvc/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java?rev=1458531&r1=1458530&r2=1458531&view=diff
==============================================================================
--- hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java (original)
+++ hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java Tue Mar 19 21:54:00 2013
@@ -41,6 +41,7 @@ import org.apache.hadoop.hbase.catalog.M
 import org.apache.hadoop.hbase.client.Put;
 import org.apache.hadoop.hbase.client.Scan;
 import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
+import org.apache.hadoop.hbase.mapreduce.hadoopbackport.JarFinder;
 import org.apache.hadoop.hbase.protobuf.ProtobufUtil;
 import org.apache.hadoop.hbase.protobuf.generated.ClientProtos;
 import org.apache.hadoop.hbase.security.User;
@@ -590,7 +591,7 @@ public class TableMapReduceUtil {
    * @return a jar file that contains the class, or null.
    * @throws IOException
    */
-  private static String findOrCreateJar(Class my_class)
+  private static String findOrCreateJar(Class<?> my_class)
   throws IOException {
     try {
       Class<?> jarFinder = Class.forName("org.apache.hadoop.util.JarFinder");
@@ -598,8 +599,7 @@ public class TableMapReduceUtil {
       // if it doesn't exist.  Note that this is needed to run the mapreduce
       // unit tests post-0.23, because mapreduce v2 requires the relevant jars
       // to be in the mr cluster to do output, split, etc.  At unit test time,
-      // the hbase jars do not exist, so we need to create some.  Note that we
-      // can safely fall back to findContainingJars for pre-0.23 mapreduce.
+      // the hbase jars do not exist, so we need to create some.
       Method m = jarFinder.getMethod("getJar", Class.class);
       return (String)m.invoke(null,my_class);
     } catch (InvocationTargetException ite) {
@@ -607,52 +607,19 @@ public class TableMapReduceUtil {
       throw new IOException(ite.getCause());
     } catch (Exception e) {
       // ignore all other exceptions. related to reflection failure
-  }
-
-  LOG.debug("New JarFinder: org.apache.hadoop.util.JarFinder.getJar " +
-	"not available.  Using old findContainingJar");
-  return findContainingJar(my_class);
-}
-
-  /**
-   * Find a jar that contains a class of the same name, if any.
-   * It will return a jar file, even if that is not the first thing
-   * on the class path that has a class with the same name.
-   *
-   * This is shamelessly copied from JobConf
-   *
-   * @param my_class the class to find.
-   * @return a jar file that contains the class, or null.
-   * @throws IOException
-   */
-  private static String findContainingJar(Class my_class) {
-    ClassLoader loader = my_class.getClassLoader();
-    String class_file = my_class.getName().replaceAll("\\.", "/") + ".class";
-    try {
-      for(Enumeration itr = loader.getResources(class_file);
-          itr.hasMoreElements();) {
-        URL url = (URL) itr.nextElement();
-        if ("jar".equals(url.getProtocol())) {
-          String toReturn = url.getPath();
-          if (toReturn.startsWith("file:")) {
-            toReturn = toReturn.substring("file:".length());
-          }
-          // URLDecoder is a misnamed class, since it actually decodes
-          // x-www-form-urlencoded MIME type rather than actual
-          // URL encoding (which the file path has). Therefore it would
-          // decode +s to ' 's which is incorrect (spaces are actually
-          // either unencoded or encoded as "%20"). Replace +s first, so
-          // that they are kept sacred during the decoding process.
-          toReturn = toReturn.replaceAll("\\+", "%2B");
-          toReturn = URLDecoder.decode(toReturn, "UTF-8");
-          return toReturn.replaceAll("!.*$", "");
-        }
-      }
-    } catch (IOException e) {
-      throw new RuntimeException(e);
     }
-    return null;
-  }
-
 
+    LOG.debug("New JarFinder: org.apache.hadoop.util.JarFinder.getJar " +
+        "not available. Falling back to backported JarFinder");
+    // Use JarFinder because it will construct a jar from class files when
+    // one does not exist. This is relevant for cases when an HBase MR job
+    // is created in the context of another MR job (particularly common for
+    // tools consuming the bulk import APIs). In that case, the dependency
+    // jars have already been shipped to and expanded in the job's working
+    // directory, so it has no jars to package. We could just construct a
+    // classpath from those class files, but we don't know the context: are
+    // they on the local filesystem, are they are ephemeral tmp files, &c.
+    // Better to package them up and ship them via the normal means.
+    return JarFinder.getJar(my_class);
+  }
 }

Added: hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/JarFinder.java
URL: http://svn.apache.org/viewvc/hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/JarFinder.java?rev=1458531&view=auto
==============================================================================
--- hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/JarFinder.java (added)
+++ hbase/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/JarFinder.java Tue Mar 19 21:54:00 2013
@@ -0,0 +1,170 @@
+/**
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License. See accompanying LICENSE file.
+ */
+package org.apache.hadoop.hbase.mapreduce.hadoopbackport;
+
+import com.google.common.base.Preconditions;
+
+import java.io.BufferedOutputStream;
+import java.io.File;
+import java.io.FileInputStream;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.net.URL;
+import java.net.URLDecoder;
+import java.text.MessageFormat;
+import java.util.Enumeration;
+import java.util.jar.JarFile;
+import java.util.jar.JarOutputStream;
+import java.util.jar.Manifest;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipOutputStream;
+
+/**
+ * Finds the Jar for a class. If the class is in a directory in the
+ * classpath, it creates a Jar on the fly with the contents of the directory
+ * and returns the path to that Jar. If a Jar is created, it is created in
+ * the system temporary directory.
+ * 
+ * This file was forked from hadoop/common/branches/branch-2@1377176.
+ */
+public class JarFinder {
+
+  private static void copyToZipStream(InputStream is, ZipEntry entry,
+                              ZipOutputStream zos) throws IOException {
+    zos.putNextEntry(entry);
+    byte[] arr = new byte[4096];
+    int read = is.read(arr);
+    while (read > -1) {
+      zos.write(arr, 0, read);
+      read = is.read(arr);
+    }
+    is.close();
+    zos.closeEntry();
+  }
+
+  public static void jarDir(File dir, String relativePath, ZipOutputStream zos)
+    throws IOException {
+    Preconditions.checkNotNull(relativePath, "relativePath");
+    Preconditions.checkNotNull(zos, "zos");
+
+    // by JAR spec, if there is a manifest, it must be the first entry in the
+    // ZIP.
+    File manifestFile = new File(dir, JarFile.MANIFEST_NAME);
+    ZipEntry manifestEntry = new ZipEntry(JarFile.MANIFEST_NAME);
+    if (!manifestFile.exists()) {
+      zos.putNextEntry(manifestEntry);
+      new Manifest().write(new BufferedOutputStream(zos));
+      zos.closeEntry();
+    } else {
+      InputStream is = new FileInputStream(manifestFile);
+      copyToZipStream(is, manifestEntry, zos);
+    }
+    zos.closeEntry();
+    zipDir(dir, relativePath, zos, true);
+    zos.close();
+  }
+
+  private static void zipDir(File dir, String relativePath, ZipOutputStream zos,
+                             boolean start) throws IOException {
+    String[] dirList = dir.list();
+    for (String aDirList : dirList) {
+      File f = new File(dir, aDirList);
+      if (!f.isHidden()) {
+        if (f.isDirectory()) {
+          if (!start) {
+            ZipEntry dirEntry = new ZipEntry(relativePath + f.getName() + "/");
+            zos.putNextEntry(dirEntry);
+            zos.closeEntry();
+          }
+          String filePath = f.getPath();
+          File file = new File(filePath);
+          zipDir(file, relativePath + f.getName() + "/", zos, false);
+        }
+        else {
+          String path = relativePath + f.getName();
+          if (!path.equals(JarFile.MANIFEST_NAME)) {
+            ZipEntry anEntry = new ZipEntry(path);
+            InputStream is = new FileInputStream(f);
+            copyToZipStream(is, anEntry, zos);
+          }
+        }
+      }
+    }
+  }
+
+  private static void createJar(File dir, File jarFile) throws IOException {
+    Preconditions.checkNotNull(dir, "dir");
+    Preconditions.checkNotNull(jarFile, "jarFile");
+    File jarDir = jarFile.getParentFile();
+    if (!jarDir.exists()) {
+      if (!jarDir.mkdirs()) {
+        throw new IOException(MessageFormat.format("could not create dir [{0}]",
+                                                   jarDir));
+      }
+    }
+    JarOutputStream zos = new JarOutputStream(new FileOutputStream(jarFile));
+    jarDir(dir, "", zos);
+  }
+
+  /**
+   * Returns the full path to the Jar containing the class. It always return a
+   * JAR.
+   *
+   * @param klass class.
+   *
+   * @return path to the Jar containing the class.
+   */
+  public static String getJar(Class klass) {
+    Preconditions.checkNotNull(klass, "klass");
+    ClassLoader loader = klass.getClassLoader();
+    if (loader != null) {
+      String class_file = klass.getName().replaceAll("\\.", "/") + ".class";
+      try {
+        for (Enumeration itr = loader.getResources(class_file);
+             itr.hasMoreElements(); ) {
+          URL url = (URL) itr.nextElement();
+          String path = url.getPath();
+          if (path.startsWith("file:")) {
+            path = path.substring("file:".length());
+          }
+          path = URLDecoder.decode(path, "UTF-8");
+          if ("jar".equals(url.getProtocol())) {
+            path = URLDecoder.decode(path, "UTF-8");
+            return path.replaceAll("!.*$", "");
+          }
+          else if ("file".equals(url.getProtocol())) {
+            String klassName = klass.getName();
+            klassName = klassName.replace(".", "/") + ".class";
+            path = path.substring(0, path.length() - klassName.length());
+            File baseDir = new File(path);
+            File testDir = new File(System.getProperty("test.build.dir", "target/test-dir"));
+            testDir = testDir.getAbsoluteFile();
+            if (!testDir.exists()) {
+              testDir.mkdirs();
+            }
+            File tempJar = File.createTempFile("hadoop-", "", testDir);
+            tempJar = new File(tempJar.getAbsolutePath() + ".jar");
+            createJar(baseDir, tempJar);
+            return tempJar.getAbsolutePath();
+          }
+        }
+      }
+      catch (IOException e) {
+        throw new RuntimeException(e);
+      }
+    }
+    return null;
+  }
+}

Added: hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/TestJarFinder.java
URL: http://svn.apache.org/viewvc/hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/TestJarFinder.java?rev=1458531&view=auto
==============================================================================
--- hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/TestJarFinder.java (added)
+++ hbase/trunk/hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/hadoopbackport/TestJarFinder.java Tue Mar 19 21:54:00 2013
@@ -0,0 +1,129 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.hbase.mapreduce.hadoopbackport;
+
+import org.apache.commons.logging.LogFactory;
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
+import java.io.File;
+import java.io.FileOutputStream;
+import java.io.FileWriter;
+import java.io.IOException;
+import java.io.OutputStream;
+import java.io.Writer;
+import java.text.MessageFormat;
+import java.util.Properties;
+import java.util.jar.JarInputStream;
+import java.util.jar.JarOutputStream;
+import java.util.jar.Manifest;
+
+/**
+ * This file was forked from hadoop/common/branches/branch-2@1350012.
+ */
+public class TestJarFinder {
+
+  @Test
+  public void testJar() throws Exception {
+
+    //picking a class that is for sure in a JAR in the classpath
+    String jar = JarFinder.getJar(LogFactory.class);
+    Assert.assertTrue(new File(jar).exists());
+  }
+
+  private static void delete(File file) throws IOException {
+    if (file.getAbsolutePath().length() < 5) {
+      throw new IllegalArgumentException(
+        MessageFormat.format("Path [{0}] is too short, not deleting",
+                             file.getAbsolutePath()));
+    }
+    if (file.exists()) {
+      if (file.isDirectory()) {
+        File[] children = file.listFiles();
+        if (children != null) {
+          for (File child : children) {
+            delete(child);
+          }
+        }
+      }
+      if (!file.delete()) {
+        throw new RuntimeException(
+          MessageFormat.format("Could not delete path [{0}]",
+                               file.getAbsolutePath()));
+      }
+    }
+  }
+
+  @Test
+  public void testExpandedClasspath() throws Exception {
+    //picking a class that is for sure in a directory in the classpath
+    //in this case the JAR is created on the fly
+    String jar = JarFinder.getJar(TestJarFinder.class);
+    Assert.assertTrue(new File(jar).exists());
+  }
+
+  @Test
+  public void testExistingManifest() throws Exception {
+    File dir = new File(System.getProperty("test.build.dir", "target/test-dir"),
+                        TestJarFinder.class.getName() + "-testExistingManifest");
+    delete(dir);
+    dir.mkdirs();
+
+    File metaInfDir = new File(dir, "META-INF");
+    metaInfDir.mkdirs();
+    File manifestFile = new File(metaInfDir, "MANIFEST.MF");
+    Manifest manifest = new Manifest();
+    OutputStream os = new FileOutputStream(manifestFile);
+    manifest.write(os);
+    os.close();
+
+    File propsFile = new File(dir, "props.properties");
+    Writer writer = new FileWriter(propsFile);
+    new Properties().store(writer, "");
+    writer.close();
+    ByteArrayOutputStream baos = new ByteArrayOutputStream();
+    JarOutputStream zos = new JarOutputStream(baos);
+    JarFinder.jarDir(dir, "", zos);
+    JarInputStream jis =
+      new JarInputStream(new ByteArrayInputStream(baos.toByteArray()));
+    Assert.assertNotNull(jis.getManifest());
+    jis.close();
+  }
+
+  @Test
+  public void testNoManifest() throws Exception {
+    File dir = new File(System.getProperty("test.build.dir", "target/test-dir"),
+                        TestJarFinder.class.getName() + "-testNoManifest");
+    delete(dir);
+    dir.mkdirs();
+    File propsFile = new File(dir, "props.properties");
+    Writer writer = new FileWriter(propsFile);
+    new Properties().store(writer, "");
+    writer.close();
+    ByteArrayOutputStream baos = new ByteArrayOutputStream();
+    JarOutputStream zos = new JarOutputStream(baos);
+    JarFinder.jarDir(dir, "", zos);
+    JarInputStream jis =
+      new JarInputStream(new ByteArrayInputStream(baos.toByteArray()));
+    Assert.assertNotNull(jis.getManifest());
+    jis.close();
+  }
+}