From db9dad4aa783d99e937ab946f3c1d0479ceb83d8 Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Thu, 2 Mar 2023 09:11:37 +0100
Subject: [PATCH] [actionmanager] increased spark.sql.shuffle.partitions for
 publication, dataset, relation records

---
 .../dhp/actionmanager/wf/dataset/oozie_app/workflow.xml       | 4 ++--
 .../dhp/actionmanager/wf/publication/oozie_app/workflow.xml   | 4 ++--
 .../dhp/actionmanager/wf/relation/oozie_app/workflow.xml      | 2 +-
 3 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/dataset/oozie_app/workflow.xml b/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/dataset/oozie_app/workflow.xml
index 4dc250c29..4f374a75a 100644
--- a/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/dataset/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/dataset/oozie_app/workflow.xml
@@ -107,7 +107,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
-                --conf spark.sql.shuffle.partitions=2560
+                --conf spark.sql.shuffle.partitions=7000
             </spark-opts>
             <arg>--inputGraphTablePath</arg><arg>${inputGraphRootPath}/dataset</arg>
             <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
@@ -159,7 +159,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
-                --conf spark.sql.shuffle.partitions=2560
+                --conf spark.sql.shuffle.partitions=7000
             </spark-opts>
             <arg>--inputGraphTablePath</arg><arg>${workingDir}/dataset</arg>
             <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
diff --git a/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/publication/oozie_app/workflow.xml b/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/publication/oozie_app/workflow.xml
index b1c8e7c85..b76dc82f1 100644
--- a/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/publication/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/publication/oozie_app/workflow.xml
@@ -107,7 +107,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
-                --conf spark.sql.shuffle.partitions=5000
+                --conf spark.sql.shuffle.partitions=7000
             </spark-opts>
             <arg>--inputGraphTablePath</arg><arg>${inputGraphRootPath}/publication</arg>
             <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
@@ -159,7 +159,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
-                --conf spark.sql.shuffle.partitions=5000
+                --conf spark.sql.shuffle.partitions=7000
             </spark-opts>
             <arg>--inputGraphTablePath</arg><arg>${workingDir}/publication</arg>
             <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
diff --git a/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/relation/oozie_app/workflow.xml b/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/relation/oozie_app/workflow.xml
index 20ffe26d3..d3086dbdc 100644
--- a/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/relation/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/relation/oozie_app/workflow.xml
@@ -99,7 +99,7 @@
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
-                --conf spark.sql.shuffle.partitions=5000
+                --conf spark.sql.shuffle.partitions=10000
             </spark-opts>
             <arg>--inputGraphTablePath</arg><arg>${inputGraphRootPath}/relation</arg>
             <arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Relation</arg>