dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml

@@ -362,7 +362,7 @@
             <arg>--inputGraphRootPath</arg><arg>${inputGraphRootPath}</arg>
             <arg>--inputRelatedEntitiesPath</arg><arg>${workingDir}/join_partial</arg>
             <arg>--outputPath</arg><arg>${workingDir}/join_entities</arg>
-            <arg>--numPartitions</arg><arg>24000</arg>
+            <arg>--numPartitions</arg><arg>35000</arg>
         </spark>
         <ok to="adjancency_lists"/>
         <error to="Kill"/>
@@ -383,7 +383,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
-                --conf spark.sql.shuffle.partitions=7680
+                --conf spark.sql.shuffle.partitions=15000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputPath</arg><arg>${workingDir}/join_entities</arg>