Merge branch 'master' into dal

addons/hive-bridge/pom.xml

@@ -90,15 +90,9 @@
 
         <dependency>
             <groupId>org.apache.hadoop</groupId>
-            <artifactId>hadoop-common</artifactId>
-            <version>${hadoop.version}</version>
-            <scope>provided</scope>
-        </dependency>
-
-        <dependency>
-            <groupId>org.apache.hadoop</groupId>
             <artifactId>hadoop-client</artifactId>
             <version>${hadoop.version}</version>
+            <scope>provided</scope>
         </dependency>
 
         <dependency>

addons/hive-bridge/src/bin/import-hive.sh

@@ -59,9 +59,11 @@ for i in "${BASEDIR}/bridge/hive/"*.jar; do
   METADATACPPATH="${METADATACPPATH}:$i"
 done
 
-echo $METADATACPPATH
+# log dir for applications
+METADATA_LOG_DIR="${METADATA_LOG_DIR:-$BASEDIR/logs}"
+export METADATA_LOG_DIR
 
-JAVA_PROPERTIES="$METADATA_OPTS"
+JAVA_PROPERTIES="$METADATA_OPTS -Dmetadata.log.dir=$METADATA_LOG_DIR -Dmetadata.log.file=import-hive.log"
 shift
 
 while [[ ${1} =~ ^\-D ]]; do
@@ -70,6 +72,7 @@ while [[ ${1} =~ ^\-D ]]; do
 done
 TIME=`date +%Y%m%d%H%M%s`
 
+#Add hive conf in classpath
 if [ ! -z "$HIVE_CONF_DIR" ]; then
     HIVE_CP=$HIVE_CONF_DIR
 elif [ ! -z "$HIVE_HOME" ]; then
@@ -86,5 +89,5 @@ echo Using Hive configuration directory [$HIVE_CP]
 ${JAVA_BIN} ${JAVA_PROPERTIES} -cp ${HIVE_CP}:${METADATACPPATH} org.apache.hadoop.metadata.hive.bridge.HiveMetaStoreBridge
 
 RETVAL=$?
-[ $RETVAL -eq 0 ] && echo Hive Data Model Imported!!!
-[ $RETVAL -ne 0 ] && echo Failure in Hive Data Model import!!!
+[ $RETVAL -eq 0 ] && echo Hive Data Model imported successfully!!!
+[ $RETVAL -ne 0 ] && echo Failed to import Hive Data Model!!!

addons/hive-bridge/src/main/java/org/apache/hadoop/metadata/hive/hook/HiveHook.java

@@ -155,7 +155,7 @@ public class HiveHook implements ExecuteWithHookContext, HiveSemanticAnalyzerHoo
 
         // clone to avoid concurrent access
         final HiveConf conf = new HiveConf(hookContext.getConf());
-        boolean debug = conf.get("debug", "false").equals("true");
+        boolean debug = conf.get("hive.hook.dgi.synchronous", "false").equals("true");
 
         if (debug) {
             fireAndForget(hookContext, conf);
@@ -178,8 +178,9 @@ public class HiveHook implements ExecuteWithHookContext, HiveSemanticAnalyzerHoo
     private void fireAndForget(HookContext hookContext, HiveConf conf) throws Exception {
         assert hookContext.getHookType() == HookContext.HookType.POST_EXEC_HOOK : "Non-POST_EXEC_HOOK not supported!";
 
+        LOG.info("Entered DGI hook for hook type {} operation {}", hookContext.getHookType(),
+                hookContext.getOperationName());
         HiveOperation operation = HiveOperation.valueOf(hookContext.getOperationName());
-        LOG.info("Entered DGI hook for hook type {} operation {}", hookContext.getHookType(), operation);
 
         HiveMetaStoreBridge dgiBridge = new HiveMetaStoreBridge(conf);
 
@@ -206,7 +207,7 @@ public class HiveHook implements ExecuteWithHookContext, HiveSemanticAnalyzerHoo
                         Table table = entity.getTable();
                         //TODO table.getDbName().toLowerCase() is required as hive stores in lowercase,
                         // but table.getDbName() is not lowercase
-                        Referenceable dbReferenceable = getDatabaseReference(dgiBridge, table.getDbName().toLowerCase());
+                        Referenceable dbReferenceable = dgiBridge.registerDatabase(table.getDbName().toLowerCase());
                         dgiBridge.registerTable(dbReferenceable, table.getDbName(), table.getTableName());
                     }
                 }
@@ -229,7 +230,8 @@ public class HiveHook implements ExecuteWithHookContext, HiveSemanticAnalyzerHoo
             LOG.info("Explain statement. Skipping...");
         }
 
-        String user = hookContext.getUserName();
+        //todo hookContext.getUserName() is null in hdp sandbox 2.2.4
+        String user = hookContext.getUserName() == null ? System.getProperty("user.name") : hookContext.getUserName();
         HiveOperation operation = HiveOperation.valueOf(hookContext.getOperationName());
         String queryId = null;
         String queryStr = null;
@@ -252,19 +254,19 @@ public class HiveHook implements ExecuteWithHookContext, HiveSemanticAnalyzerHoo
             if (readEntity.getTyp() == Entity.Type.TABLE) {
                 Table table = readEntity.getTable();
                 String dbName = table.getDbName().toLowerCase();
-                source.add(getTableReference(dgiBridge, dbName, table.getTableName()));
+                source.add(dgiBridge.registerTable(dbName, table.getTableName()));
             }
         }
-        processReferenceable.set("sourceTableNames", source);
+        processReferenceable.set("inputTables", source);
         List<Referenceable> target = new ArrayList<>();
         for (WriteEntity writeEntity : outputs) {
             if (writeEntity.getTyp() == Entity.Type.TABLE) {
                 Table table = writeEntity.getTable();
                 String dbName = table.getDbName().toLowerCase();
-                target.add(getTableReference(dgiBridge, dbName, table.getTableName()));
+                target.add(dgiBridge.registerTable(dbName, table.getTableName()));
             }
         }
-        processReferenceable.set("targetTableNames", target);
+        processReferenceable.set("outputTables", target);
         processReferenceable.set("queryText", queryStr);
         processReferenceable.set("queryId", queryId);
         processReferenceable.set("queryPlan", getQueryPlan(hookContext, conf));
@@ -275,58 +277,6 @@ public class HiveHook implements ExecuteWithHookContext, HiveSemanticAnalyzerHoo
         dgiBridge.createInstance(processReferenceable);
     }
 
-    /**
-     * Gets reference for the database. Creates new instance if it doesn't exist
-     *
-     * @param dgiBridge
-     * @param dbName    database name
-     * @return Reference for database
-     * @throws Exception
-     */
-    private Referenceable getDatabaseReference(HiveMetaStoreBridge dgiBridge, String dbName) throws Exception {
-        String typeName = HiveDataTypes.HIVE_DB.getName();
-        MetadataServiceClient dgiClient = dgiBridge.getMetadataServiceClient();
-
-        JSONObject result = dgiClient.rawSearch(typeName, "name", dbName);
-        JSONArray results = (JSONArray) result.get("results");
-
-        if (results.length() == 0) {
-            //Create new instance
-            return dgiBridge.registerDatabase(dbName);
-
-        } else {
-            String guid = (String) ((JSONObject) results.get(0)).get("guid");
-            return new Referenceable(guid, typeName, null);
-        }
-    }
-
-    /**
-     * Gets reference for the table. Creates new instance if it doesn't exist
-     *
-     * @param dgiBridge
-     * @param dbName
-     * @param tableName table name
-     * @return table reference
-     * @throws Exception
-     */
-    private Referenceable getTableReference(HiveMetaStoreBridge dgiBridge, String dbName, String tableName) throws Exception {
-        String typeName = HiveDataTypes.HIVE_TABLE.getName();
-        MetadataServiceClient dgiClient = dgiBridge.getMetadataServiceClient();
-
-        JSONObject result = dgiClient.rawSearch(typeName, "tableName", tableName);
-        JSONArray results = (JSONArray) result.get("results");
-
-        if (results.length() == 0) {
-            Referenceable dbRererence = getDatabaseReference(dgiBridge, dbName);
-            return dgiBridge.registerTable(dbRererence, dbName, tableName).first;
-
-        } else {
-            //There should be just one instance with the given name
-            String guid = (String) ((JSONObject) results.get(0)).get("guid");
-            return new Referenceable(guid, typeName, null);
-        }
-    }
-
 
     private String getQueryPlan(HookContext hookContext, HiveConf conf) throws Exception {
         //We need to somehow get the sem associated with the plan and use it here.

addons/hive-bridge/src/main/java/org/apache/hadoop/metadata/hive/model/HiveDataModelGenerator.java

@@ -340,8 +340,8 @@ public class HiveDataModelGenerator {
     private void createPartitionClass() throws MetadataException {
 
         AttributeDefinition[] attributeDefinitions = new AttributeDefinition[]{
-                new AttributeDefinition("values", DataTypes.STRING_TYPE.getName(),
-                        Multiplicity.COLLECTION, false, null),
+                new AttributeDefinition("values", DataTypes.arrayTypeName(DataTypes.STRING_TYPE.getName()),
+                        Multiplicity.OPTIONAL, false, null),
                 new AttributeDefinition("dbName", HiveDataTypes.HIVE_DB.getName(),
                         Multiplicity.REQUIRED, false, null),
                 new AttributeDefinition("tableName", HiveDataTypes.HIVE_TABLE.getName(),
@@ -354,10 +354,9 @@ public class HiveDataModelGenerator {
                         Multiplicity.REQUIRED, false, null),
                 new AttributeDefinition("columns",
                         DataTypes.arrayTypeName(HiveDataTypes.HIVE_COLUMN.getName()),
-                        Multiplicity.COLLECTION, true, null),
+                        Multiplicity.OPTIONAL, true, null),
                 new AttributeDefinition("parameters", STRING_MAP_TYPE.getName(),
                         Multiplicity.OPTIONAL, false, null),
-
         };
         HierarchicalTypeDefinition<ClassType> definition =
                 new HierarchicalTypeDefinition<>(ClassType.class,

addons/hive-bridge/src/site/twiki/Bridge-Hive.twiki

@@ -21,8 +21,8 @@ Hive metadata can be modelled in DGI using its Type System. The default modellin
 
 ---++ Importing Hive Metadata
 org.apache.hadoop.metadata.hive.bridge.HiveMetaStoreBridge imports the hive metadata into DGI using the typesystem defined in org.apache.hadoop.metadata.hive.model.HiveDataModelGenerator. import-hive.sh command can be used to facilitate this.
-Set-up the following configs in <dgi package>/conf/hive-site.xml:
-   * Hive metastore configuration - Refer [[https://cwiki.apache.org/confluence/display/Hive/AdminManual+MetastoreAdmin][Hive Metastore Configuration documentation]]
+Set-up the following configs in hive-site.xml of your hive set-up and set environment variable HIVE_CONFIG to the
+hive conf directory:
    * DGI endpoint - Add the following property with the DGI endpoint for your set-up
 <verbatim>
 <property>
@@ -57,4 +57,5 @@ The following properties in hive-site.xml control the thread pool details:
    * hive.hook.dgi.minThreads - core number of threads. default 5
    * hive.hook.dgi.maxThreads - maximum number of threads. default 5
    * hive.hook.dgi.keepAliveTime - keep alive time in msecs. default 10
+   * hive.hook.dgi.synchronous - boolean, true to run the hook synchronously. default false
 

addons/hive-bridge/src/test/java/org/apache/hadoop/metadata/hive/hook/HiveHookIT.java

@@ -58,7 +58,7 @@ public class HiveHookIT {
         hiveConf.setVar(HiveConf.ConfVars.METASTOREWAREHOUSE, System.getProperty("user.dir") + "/target/metastore");
         hiveConf.set(HiveMetaStoreBridge.DGI_URL_PROPERTY, DGI_URL);
         hiveConf.set("javax.jdo.option.ConnectionURL", "jdbc:derby:./target/metastore_db;create=true");
-        hiveConf.set("debug", "true");
+        hiveConf.set("hive.hook.dgi.synchronous", "true");
         return hiveConf;
     }
 
@@ -114,10 +114,7 @@ public class HiveHookIT {
     }
 
     private void assertInstanceIsRegistered(String typeName, String colName, String colValue) throws Exception{
-        JSONObject result = dgiCLient.rawSearch(typeName, colName, colValue);
-        JSONArray results = (JSONArray) result.get("results");
+        JSONArray results = dgiCLient.rawSearch(typeName, colName, colValue);
         Assert.assertEquals(results.length(), 1);
-        JSONObject resultRow = (JSONObject) results.get(0);
-        Assert.assertEquals(resultRow.get(typeName + "." + colName), colValue);
     }
 }

client/src/main/java/org/apache/hadoop/metadata/MetadataServiceClient.java

@@ -22,6 +22,10 @@ import com.sun.jersey.api.client.Client;
 import com.sun.jersey.api.client.ClientResponse;
 import com.sun.jersey.api.client.WebResource;
 import com.sun.jersey.api.client.config.DefaultClientConfig;
+import org.apache.hadoop.metadata.typesystem.ITypedReferenceableInstance;
+import org.apache.hadoop.metadata.typesystem.Referenceable;
+import org.apache.hadoop.metadata.typesystem.json.InstanceSerialization;
+import org.apache.hadoop.metadata.typesystem.json.Serialization;
 import org.codehaus.jettison.json.JSONArray;
 import org.codehaus.jettison.json.JSONException;
 import org.codehaus.jettison.json.JSONObject;
@@ -149,8 +153,14 @@ public class MetadataServiceClient {
      * @return result json object
      * @throws MetadataServiceException
      */
-    public JSONObject getEntity(String guid) throws MetadataServiceException {
-        return callAPI(API.GET_ENTITY, null, guid);
+    public ITypedReferenceableInstance getEntity(String guid) throws MetadataServiceException {
+        JSONObject jsonResponse = callAPI(API.GET_ENTITY, null, guid);
+        try {
+            String entityInstanceDefinition = jsonResponse.getString(MetadataServiceClient.RESULTS);
+            return Serialization.fromJson(entityInstanceDefinition);
+        } catch (JSONException e) {
+            throw new MetadataServiceException(e);
+        }
     }
 
     public JSONObject searchEntity(String searchQuery) throws MetadataServiceException {
@@ -167,14 +177,14 @@ public class MetadataServiceClient {
      * @return result json object
      * @throws MetadataServiceException
      */
-    public JSONObject rawSearch(String typeName, String attributeName,
-                                Object attributeValue) throws MetadataServiceException {
-        String gremlinQuery = String.format(
-                "g.V.has(\"typeName\",\"%s\").and(_().has(\"%s.%s\", T.eq, \"%s\")).toList()",
-                typeName, typeName, attributeName, attributeValue);
-        return searchByGremlin(gremlinQuery);
-//        String dslQuery = String.format("%s where %s = \"%s\"", typeName, attributeName, attributeValue);
-//        return searchByDSL(dslQuery);
+    public JSONArray rawSearch(String typeName, String attributeName, Object attributeValue) throws
+            MetadataServiceException {
+//        String gremlinQuery = String.format(
+//                "g.V.has(\"typeName\",\"%s\").and(_().has(\"%s.%s\", T.eq, \"%s\")).toList()",
+//                typeName, typeName, attributeName, attributeValue);
+//        return searchByGremlin(gremlinQuery);
+        String dslQuery = String.format("%s where %s = \"%s\"", typeName, attributeName, attributeValue);
+        return searchByDSL(dslQuery);
     }
 
     /**
@@ -183,10 +193,15 @@ public class MetadataServiceClient {
      * @return result json object
      * @throws MetadataServiceException
      */
-    public JSONObject searchByDSL(String query) throws MetadataServiceException {
+    public JSONArray searchByDSL(String query) throws MetadataServiceException {
         WebResource resource = getResource(API.SEARCH_DSL);
         resource = resource.queryParam("query", query);
-        return callAPIWithResource(API.SEARCH_DSL, resource);
+        JSONObject result = callAPIWithResource(API.SEARCH_DSL, resource);
+        try {
+            return result.getJSONObject("results").getJSONArray("rows");
+        } catch (JSONException e) {
+            throw new MetadataServiceException(e);
+        }
     }
 
     /**

repository/src/main/scala/org/apache/hadoop/metadata/query/GremlinEvaluator.scala

@@ -104,7 +104,6 @@ class GremlinEvaluator(qry: GremlinQuery, persistenceStrategy: GraphPersistenceS
                     val v = rV.getColumn(src).get(idx)
                     sInstance.set(cName, persistenceStrategy.constructInstance(aE.dataType, v))
                 }
-                sInstance
               addPathStruct(r, sInstance)
             }
             GremlinQueryResult(qry.expr.toString, rType, rows.toList)

src/bin/metadata-config.sh

@@ -99,7 +99,7 @@ mkdir -p $METADATA_LOG_DIR
 
 pushd ${BASEDIR} > /dev/null
 
-JAVA_PROPERTIES="$METADATA_OPTS $METADATA_PROPERTIES -Dmetadata.log.dir=$METADATA_LOG_DIR -Dmetadata.home=${METADATA_HOME_DIR} -Dmetadata.conf=${METADATA_CONF}"
+JAVA_PROPERTIES="$METADATA_OPTS $METADATA_PROPERTIES -Dmetadata.log.dir=$METADATA_LOG_DIR -Dmetadata.home=${METADATA_HOME_DIR} -Dmetadata.conf=${METADATA_CONF} -Dmetadata.log.file=application.log"
 shift
 
 while [[ ${1} =~ ^\-D ]]; do

src/conf/log4j.xml

@@ -28,7 +28,7 @@
     </appender>
 
     <appender name="FILE" class="org.apache.log4j.DailyRollingFileAppender">
-        <param name="File" value="${metadata.log.dir}/application.log"/>
+        <param name="File" value="${metadata.log.dir}/${metadata.log.file}"/>
         <param name="Append" value="true"/>
         <param name="Threshold" value="debug"/>
         <layout class="org.apache.log4j.PatternLayout">