From b39ff36c1641c647368006a35e4bfb278c75ab26 Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Thu, 16 Apr 2020 12:11:37 +0200
Subject: [PATCH] improving the wf definitions

---
 .../dhp/oa/dedup/scan/oozie_app/workflow.xml  | 67 ++++++++++++-------
 1 file changed, 44 insertions(+), 23 deletions(-)

diff --git a/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml b/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml
index 25596bc2f..b50241b8f 100644
--- a/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml
@@ -28,6 +28,28 @@
             <name>sparkExecutorCores</name>
             <description>number of cores used by single executor</description>
         </property>
+        <property>
+            <name>oozieActionShareLibForSpark2</name>
+            <description>oozie action sharelib for spark 2.*</description>
+        </property>
+        <property>
+            <name>spark2ExtraListeners</name>
+            <value>com.cloudera.spark.lineage.NavigatorAppListener</value>
+            <description>spark 2.* extra listeners classname</description>
+        </property>
+        <property>
+            <name>spark2SqlQueryExecutionListeners</name>
+            <value>com.cloudera.spark.lineage.NavigatorQueryListener</value>
+            <description>spark 2.* sql query execution listeners classname</description>
+        </property>
+        <property>
+            <name>spark2YarnHistoryServerAddress</name>
+            <description>spark 2.* yarn history server address</description>
+        </property>
+        <property>
+            <name>spark2EventLogDir</name>
+            <description>spark 2.* event log dir location</description>
+        </property>
     </parameters>
 
     <global>
@@ -42,6 +64,10 @@
                 <name>oozie.launcher.mapred.job.queue.name</name>
                 <value>${oozieLauncherQueueName}</value>
             </property>
+            <property>
+                <name>oozie.action.sharelib.for.spark</name>
+                <value>${oozieActionShareLibForSpark2}</value>
+            </property>
         </configuration>
     </global>
 
@@ -53,20 +79,19 @@
 
     <action name="CreateSimRel">
         <spark xmlns="uri:oozie:spark-action:0.2">
-            <prepare>
-                <delete path="${workingPath}/${actionSetId}/*_simrel"/>
-            </prepare>
             <master>yarn</master>
             <mode>cluster</mode>
             <name>Create Similarity Relations</name>
             <class>eu.dnetlib.dhp.oa.dedup.SparkCreateSimRels</class>
             <jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
             <spark-opts>
-                --executor-memory ${sparkExecutorMemory}
-                --executor-cores ${sparkExecutorCores}
+                --executor-memory=${sparkExecutorMemory}
+                --executor-cores=${sparkExecutorCores}
                 --driver-memory=${sparkDriverMemory}
-                --conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
-                --conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
+                --conf spark.extraListeners=${spark2ExtraListeners}
+                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
+                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
+                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
             </spark-opts>
             <arg>-mt</arg><arg>yarn</arg>
             <arg>--i</arg><arg>${graphBasePath}</arg>
@@ -80,21 +105,19 @@
 
     <action name="CreateMergeRel">
         <spark xmlns="uri:oozie:spark-action:0.2">
-            <prepare>
-                <delete path='${workingPath}/${actionSetId}/*_mergerel'/>
-            </prepare>
             <master>yarn</master>
             <mode>cluster</mode>
             <name>Create Merge Relations</name>
             <class>eu.dnetlib.dhp.oa.dedup.SparkCreateConnectedComponent</class>
             <jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
             <spark-opts>
-                --executor-memory ${sparkExecutorMemory}
-                --executor-cores ${sparkExecutorCores}
+                --executor-memory=${sparkExecutorMemory}
+                --executor-cores=${sparkExecutorCores}
                 --driver-memory=${sparkDriverMemory}
-                --conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
-                --conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
-                --conf spark.sql.warehouse.dir="/user/hive/warehouse"
+                --conf spark.extraListeners=${spark2ExtraListeners}
+                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
+                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
+                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
             </spark-opts>
             <arg>-mt</arg><arg>yarn</arg>
             <arg>--i</arg><arg>${graphBasePath}</arg>
@@ -108,21 +131,19 @@
 
     <action name="CreateDedupRecord">
         <spark xmlns="uri:oozie:spark-action:0.2">
-            <prepare>
-                <delete path='${workingPath}/${actionSetId}/*_deduprecord'/>
-            </prepare>
             <master>yarn</master>
             <mode>cluster</mode>
             <name>Create Dedup Record</name>
             <class>eu.dnetlib.dhp.oa.dedup.SparkCreateDedupRecord</class>
             <jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
             <spark-opts>
-                --executor-memory ${sparkExecutorMemory}
-                --executor-cores ${sparkExecutorCores}
+                --executor-memory=${sparkExecutorMemory}
+                --executor-cores=${sparkExecutorCores}
                 --driver-memory=${sparkDriverMemory}
-                --conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
-                --conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
-                --conf spark.sql.warehouse.dir="/user/hive/warehouse"
+                --conf spark.extraListeners=${spark2ExtraListeners}
+                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
+                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
+                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
             </spark-opts>
             <arg>-mt</arg><arg>yarn</arg>
             <arg>--i</arg><arg>${graphBasePath}</arg>