From ec5dfc068db477fbbb8e335b52a6a06f8517d125 Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Thu, 16 Apr 2020 14:41:28 +0200
Subject: [PATCH] added spark.sql.shuffle.partitions=3840 to dedup scan wf

---
 .../eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml         | 2 ++
 1 file changed, 2 insertions(+)

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 250f7398fe..6c74a204dc 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
@@ -92,6 +92,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
+                --conf spark.sql.shuffle.partitions=3840
             </spark-opts>
             <arg>-mt</arg><arg>yarn</arg>
             <arg>--i</arg><arg>${graphBasePath}</arg>
@@ -118,6 +119,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
+                --conf spark.sql.shuffle.partitions=3840
             </spark-opts>
             <arg>-mt</arg><arg>yarn</arg>
             <arg>--i</arg><arg>${graphBasePath}</arg>