SQOOP-397. Make Sqoop work with Hadoop 0.23

(Bilung Lee via Arvind Prabhakar)


git-svn-id: https://svn.apache.org/repos/asf/incubator/sqoop/trunk@1211583 13f79535-47bb-0310-9956-ffa450edef68

COMPILING.txt

@@ -55,9 +55,9 @@ If you want to build everything (including the documentation), type
 +ant package+. This will appear in the
 +build/sqoop-(version)/+ directory.
 
-Sqoop is built against the latest Hadoop distribution available from Cloudera.
-These dependencies are obtained via IVY which downloads the necessary binaries
-from Cloudera maven repository.
+This version of Sqoop is built against Hadoop 0.23 available from Apache
+maven repository by default. These dependencies are obtained via IVY which
+downloads the necessary binaries.
 
 == Testing Sqoop
 
@@ -274,3 +274,12 @@ will allow you to edit Sqoop sources in Eclipse with all the library
 dependencies correctly resolved. To compile the jars, you should still
 use ant.
 
+
+== Using a specific version of Hadoop
+
+Now Sqoop defaults to use Hadoop 0.23 available from Apache maven repository.
+To switch back to the previous version of Hadoop 0.20, for example, run:
+
+++++
+ant test -Dhadoopversion=20
+++++

build.xml

@@ -24,6 +24,13 @@
     xmlns:artifact="urn:maven-artifact-ant"
     xmlns:ivy="antlib:org.apache.ivy.ant">
 
+  <!-- load ant-contrib tasks to get the "if" task. -->
+  <taskdef resource="net/sf/antcontrib/antcontrib.properties">
+    <classpath>
+      <pathelement location="${basedir}/lib/ant-contrib-1.0b3.jar"/>
+    </classpath>
+  </taskdef>
+
   <!-- Load system-wide and project-wide default properties set by
        the user, to avoid needing to override with -D. -->
   <property file="${user.home}/build.properties" />
@@ -150,10 +157,22 @@
     -->
   <property name="sqoop.thirdparty.lib.dir" value="" />
 
+  <!-- Set default Hadoop version if not set -->
+  <if>
+    <isset property="hadoopversion" />
+    <then>
+      <echo message="Use Hadoop 0.${hadoopversion}" />
+    </then>
+    <else>
+      <echo message="Use Hadoop 0.23 by default" />
+      <property name="hadoopversion" value="23" />
+    </else>
+  </if>
+
   <!-- Ivy-based dependency resolution -->
   <property name="ivy.dir" location="${basedir}/ivy" />
-  <property name="ivysettings.xml" location="${ivy.dir}/ivysettings.xml"/>
   <loadproperties srcfile="${ivy.dir}/libraries.properties"/>
+  <property name="ivysettings.xml" location="${ivy.dir}/ivysettings.xml"/>
   <property name="ivy.jar" location="${lib.dir}/ivy-${ivy.version}.jar"/>
   <property name="ivy_repo_url"
   value="http://repo2.maven.org/maven2/org/apache/ivy/ivy/${ivy.version}/ivy-${ivy.version}.jar" />
@@ -184,13 +203,6 @@
   <property name="sqoop.test.sqlserver.connectstring.host_url"
             value="jdbc:sqlserver://sqlserverhost:1433"/>
 
-  <!-- load ant-contrib tasks to get the "if" task. -->
-  <taskdef resource="net/sf/antcontrib/antcontrib.properties">
-    <classpath>
-      <pathelement location="${basedir}/lib/ant-contrib-1.0b3.jar"/>
-    </classpath>
-  </taskdef>
-
   <!-- The classpath for compiling and running Sqoop -->
   <if>
     <isset property="hadoop.home" />
@@ -1062,20 +1074,20 @@
   <!-- retrieve ivy-managed artifacts from the Hadoop distribution -->
   <target name="ivy-resolve-hadoop" depends="ivy-init"
       unless="hadoop.is.local">
-    <ivy:resolve settingsRef="${name}.ivy.settings" conf="cloudera" />
+    <ivy:resolve settingsRef="${name}.ivy.settings" conf="hadoop${hadoopversion}" />
   </target>
   <target name="ivy-retrieve-hadoop" depends="ivy-init,ivy-resolve-hadoop">
     <!-- retrieve hadoop refs normally. -->
     <ivy:retrieve settingsRef="${name}.ivy.settings"
        pattern="${build.ivy.lib.dir}/${ivy.artifact.retrieve.pattern}"
        sync="true" />
-    <ivy:cachepath pathid="${name}.hadoop.classpath" conf="cloudera" />
+    <ivy:cachepath pathid="${name}.hadoop.classpath" conf="hadoop${hadoopversion}" />
   </target>
 
   <!-- retrieve ivy-managed test artifacts from the Hadoop distribution -->
   <target name="ivy-resolve-hadoop-test" depends="ivy-init"
       unless="hadoop.is.local">
-    <ivy:resolve settingsRef="${name}.ivy.settings" conf="clouderatest" />
+    <ivy:resolve settingsRef="${name}.ivy.settings" conf="hadoop${hadoopversion}test" />
   </target>
   <target name="ivy-retrieve-hadoop-test"
       depends="ivy-init,ivy-resolve-hadoop-test">
@@ -1083,7 +1095,7 @@
           pattern="${build.ivy.lib.dir}/${ivy.artifact.retrieve.pattern}"
           sync="true" />
         <ivy:cachepath pathid="${name}.hadooptest.classpath"
-          conf="clouderatest" />
+          conf="hadoop${hadoopversion}test" />
   </target>
 
   <!-- retrieve ivy-managed artifacts for checkstyle -->

ivy.xml

@@ -36,14 +36,13 @@ under the License.
     <conf name="common" visibility="private"
       extends="runtime"
       description="artifacts needed to compile/test the application"/>
-    <conf name="cloudera" visibility="private"
-      extends="common,runtime"
-      description="artifacts from Cloudera for compile/test" />
+    <conf name="hbase" visibility="private" />
+    <conf name="hadoop23" visibility="private" extends="common,runtime,hbase" />
+    <conf name="hadoop20" visibility="private" extends="common,runtime,hbase" />
 
     <conf name="test" visibility="private" extends="common,runtime"/>
-    <conf name="clouderatest" visibility="private"
-      extends="test,cloudera"
-      description="artifacts from Cloudera for testing" />
+    <conf name="hadoop23test" visibility="private" extends="test,hadoop23" />
+    <conf name="hadoop20test" visibility="private" extends="test,hadoop20" />
 
     <!-- We don't redistribute everything we depend on (e.g., Hadoop itself);
          anything which Hadoop itself also depends on, we do not ship.
@@ -67,11 +66,28 @@ under the License.
     <artifact conf="master"/>
   </publications>
   <dependencies>
-    <!-- Dependencies for Cloudera's Distribution for Hadoop -->
+
+    <!-- Dependencies for Hadoop 0.23 -->
+    <dependency org="org.apache.hadoop" name="hadoop-common"
+      rev="${hadoop-common.version}" conf="hadoop23->default">
+      <artifact name="hadoop-common" type="jar" />
+      <artifact name="hadoop-common" type="jar" m:classifier="tests"/>
+    </dependency>
+    <dependency org="org.apache.hadoop" name="hadoop-hdfs"
+      rev="${hadoop-common.version}" conf="hadoop23->default">
+      <artifact name="hadoop-hdfs" type="jar" />
+      <artifact name="hadoop-hdfs" type="jar" m:classifier="tests"/>
+    </dependency>
+    <dependency org="org.apache.hadoop" name="hadoop-mapreduce-client-common"
+      rev="${hadoop-common.version}" conf="hadoop23->default"/>
+    <dependency org="org.apache.hadoop" name="hadoop-mapreduce-client-core"
+      rev="${hadoop-common.version}" conf="hadoop23->default"/>
+
+    <!-- Dependencies for Hadoop 0.20 -->
     <dependency org="org.apache.hadoop" name="hadoop-core"
-      rev="${hadoop-core.cloudera.version}" conf="cloudera->default"/>
+      rev="${hadoop-core.version}" conf="hadoop20->default"/>
     <dependency org="org.apache.hadoop" name="hadoop-test"
-      rev="${hadoop-core.cloudera.version}" conf="clouderatest->default"/>
+      rev="${hadoop-core.version}" conf="hadoop20test->default"/>
 
     <!-- Common dependencies for Sqoop -->
     <dependency org="commons-cli" name="commons-cli"
@@ -113,7 +129,7 @@ under the License.
       rev="${commons-collections.version}" conf="releaseaudit->default"/>
 
     <dependency org="org.apache.hbase" name="hbase"
-      rev="${hbase.version}" conf="cloudera->default">
+      rev="${hbase.version}" conf="hbase->default">
       <artifact name="hbase" type="jar"/>
       <artifact name="hbase" type="test-jar" ext="jar" m:classifier="tests"/>
       <exclude org="com.sun.jersey" module="jersey-core"/>
@@ -121,9 +137,12 @@ under the License.
       <exclude org="com.sun.jersey" module="jersey-server"/>
       <exclude org="org.apache.thrift" module="thrift"/>
       <exclude org="log4j" module="log4j"/>
+      <exclude org="org.apache.hadoop" module="hadoop-core" />
+      <exclude org="com.cloudera.cdh" module="zookeeper-ant" />
     </dependency>
 
     <exclude org="org.apache.hadoop" module="avro"/>
+    <exclude org="commons-daemon" module="commons-daemon" />
 
   </dependencies>
 </ivy-module>

ivy/libraries.properties

@@ -28,15 +28,14 @@ commons-io.version=1.4
 commons-lang.version=2.4
 commons-logging.version=1.0.4
 
-# Cloudera Distribution dependency version
-hadoop-core.cloudera.version=0.20.2-cdh3u1
-
+hadoop-core.version=0.20.2-cdh3u1
+hadoop-common.version=0.23.0-SNAPSHOT
 hbase.version=0.90.3-cdh3u1
 zookeeper.version=3.3.3-cdh3u1
 
 hsqldb.version=1.8.0.10
 
-ivy.version=2.0.0-rc2
+ivy.version=2.1.0
 
 junit.version=4.5
 

src/test/com/cloudera/sqoop/TestCompression.java

@@ -163,6 +163,7 @@ public void runTextCompressionTest(CompressionCodec codec, int expectedNum)
 
     if (codec == null) {
       codec = new GzipCodec();
+      ReflectionUtils.setConf(codec, getConf());
     }
     Path p = new Path(getDataFilePath().toString()
         + codec.getDefaultExtension());

src/test/com/cloudera/sqoop/hbase/HBaseTestCase.java

@@ -34,6 +34,7 @@
 import org.apache.hadoop.hbase.util.Bytes;
 
 import org.apache.hadoop.util.StringUtils;
+import org.apache.hadoop.util.VersionInfo;
 
 import org.junit.After;
 import org.junit.Before;
@@ -120,6 +121,9 @@ private void startMaster() throws Exception {
   @Override
   @Before
   public void setUp() {
+    if (!isHadoop20()) {
+      return;
+    }
     HBaseTestCase.recordTestBuildDataProperty();
     try {
       startMaster();
@@ -143,6 +147,9 @@ public void shutdown() throws Exception {
   @Override
   @After
   public void tearDown() {
+    if (!isHadoop20()) {
+      return;
+    }
     try {
       shutdown();
     } catch (Exception e) {
@@ -173,4 +180,8 @@ protected void verifyHBaseCell(String tableName, String rowKey,
       table.close();
     }
   }
+
+  protected boolean isHadoop20() {
+    return VersionInfo.getVersion().startsWith("0.20");
+  }
 }

src/test/com/cloudera/sqoop/hbase/TestHBaseImport.java

@@ -29,6 +29,9 @@ public class TestHBaseImport extends HBaseTestCase {
 
   @Test
   public void testBasicUsage() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     // Create the HBase table in Sqoop as we run the job.
     String [] argv = getArgv(true, "BasicUsage", "BasicColFam", true, null);
     String [] types = { "INT", "INT" };
@@ -40,6 +43,9 @@ public void testBasicUsage() throws IOException {
 
   @Test
   public void testMissingTableFails() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     // Test that if the table doesn't exist, we fail unless we
     // explicitly create the table.
     String [] argv = getArgv(true, "MissingTable", "MissingFam", false, null);
@@ -56,6 +62,9 @@ public void testMissingTableFails() throws IOException {
 
   @Test
   public void testOverwriteSucceeds() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     // Test that we can create a table and then import immediately
     // back on top of it without problem.
     String [] argv = getArgv(true, "OverwriteT", "OverwriteF", true, null);
@@ -71,6 +80,9 @@ public void testOverwriteSucceeds() throws IOException {
 
   @Test
   public void testStrings() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     String [] argv = getArgv(true, "stringT", "stringF", true, null);
     String [] types = { "INT", "VARCHAR(32)" };
     String [] vals = { "0", "'abc'" };
@@ -81,6 +93,9 @@ public void testStrings() throws IOException {
 
   @Test
   public void testNulls() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     String [] argv = getArgv(true, "nullT", "nullF", true, null);
     String [] types = { "INT", "INT", "INT" };
     String [] vals = { "0", "42", "null" };
@@ -96,6 +111,9 @@ public void testNulls() throws IOException {
 
   @Test
   public void testExitFailure() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     String [] types = { "INT", "INT", "INT" };
     String [] vals = { "0", "42", "43" };
     createTableWithColTypes(types, vals);

src/test/com/cloudera/sqoop/hbase/TestHBaseQueryImport.java

@@ -29,6 +29,9 @@ public class TestHBaseQueryImport extends HBaseTestCase {
 
   @Test
   public void testImportFromQuery() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     String [] types = { "INT", "INT", "INT" };
     String [] vals = { "0", "42", "43" };
     createTableWithColTypes(types, vals);
@@ -47,6 +50,9 @@ public void testImportFromQuery() throws IOException {
 
   @Test
   public void testExitFailure() throws IOException {
+    if (!isHadoop20()) {
+      return;
+    }
     String [] types = { "INT", "INT", "INT" };
     String [] vals = { "0", "42", "43" };
     createTableWithColTypes(types, vals);

src/test/com/cloudera/sqoop/mapreduce/db/TestDataDrivenDBInputFormat.java

@@ -35,6 +35,7 @@
 import org.apache.hadoop.mapreduce.lib.db.*;
 import org.apache.hadoop.mapreduce.lib.output.*;
 import org.apache.hadoop.util.StringUtils;
+import org.apache.hadoop.util.VersionInfo;
 
 /**
  * Test aspects of DataDrivenDBInputFormat.
@@ -82,11 +83,17 @@ private void initialize(String driverClassName, String url)
   }
 
   public void setUp() throws Exception {
+    if (!isHadoop20()) {
+      return;
+    }
     initialize(DRIVER_CLASS, DB_URL);
     super.setUp();
   }
 
   public void tearDown() throws Exception {
+    if (!isHadoop20()) {
+      return;
+    }
     super.tearDown();
     shutdown();
   }
@@ -165,6 +172,9 @@ public void map(Object k, Object v, Context c)
   }
 
   public void testDateSplits() throws Exception {
+    if (!isHadoop20()) {
+      return;
+    }
     Statement s = connection.createStatement();
     final String DATE_TABLE = "datetable";
     final String COL = "foo";
@@ -219,4 +229,8 @@ public void testDateSplits() throws Exception {
       s.close();
     }
   }
+
+  protected boolean isHadoop20() {
+    return VersionInfo.getVersion().startsWith("0.20");
+  }
 }