From cadb5a42c2fb637c869c79eaecdc9d1bda1679db Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Thu, 15 Feb 2024 14:11:25 +0100
Subject: [PATCH] removed spark.shuffle.sort.bypassMergeThreshold and increased
 spark.yarn.executor.memoryOverhead to 6G

---
 .../dhp/oa/provision/oozie_app/workflow.xml   | 32 +++++++++----------
 1 file changed, 16 insertions(+), 16 deletions(-)

diff --git a/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml b/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml
index 7948c7198..2a4449693 100644
--- a/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml
@@ -147,12 +147,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=3840
-                --conf spark.shuffle.sort.bypassMergeThreshold=3840
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${inputGraphRootPath}/relation</arg>
             <arg>--outputPath</arg><arg>${workingDir}/relation</arg>
@@ -186,12 +186,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=15000
-                --conf spark.shuffle.sort.bypassMergeThreshold=15000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -214,12 +214,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=15000
-                --conf spark.shuffle.sort.bypassMergeThreshold=15000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -242,12 +242,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=10000
-                --conf spark.shuffle.sort.bypassMergeThreshold=10000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -270,12 +270,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --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.shuffle.sort.bypassMergeThreshold=5000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -298,12 +298,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --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.shuffle.sort.bypassMergeThreshold=5000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -326,12 +326,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --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.shuffle.sort.bypassMergeThreshold=5000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -354,12 +354,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --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.shuffle.sort.bypassMergeThreshold=5000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${workingDir}/relation</arg>
@@ -394,12 +394,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=15000
-                --conf spark.shuffle.sort.bypassMergeThreshold=15000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/publication</arg>
@@ -423,12 +423,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=10000
-                --conf spark.shuffle.sort.bypassMergeThreshold=10000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/dataset</arg>
@@ -452,12 +452,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=10000
-                --conf spark.shuffle.sort.bypassMergeThreshold=10000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/otherresearchproduct</arg>
@@ -481,12 +481,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --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.shuffle.sort.bypassMergeThreshold=5000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/software</arg>
@@ -510,12 +510,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=8000
-                --conf spark.shuffle.sort.bypassMergeThreshold=8000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/datasource</arg>
@@ -539,12 +539,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=10000
-                --conf spark.shuffle.sort.bypassMergeThreshold=10000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/organization</arg>
@@ -568,12 +568,12 @@
                 --executor-cores=${sparkExecutorCoresForJoining}
                 --executor-memory=${sparkExecutorMemoryForJoining}
                 --driver-memory=${sparkDriverMemoryForJoining}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --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.shuffle.sort.bypassMergeThreshold=5000
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputEntityPath</arg><arg>${inputGraphRootPath}/project</arg>
@@ -599,12 +599,12 @@
                 --executor-cores=${sparkExecutorCores}
                 --executor-memory=${sparkExecutorMemory}
                 --driver-memory=${sparkDriverMemory}
+                --conf spark.yarn.executor.memoryOverhead=6G
                 --conf spark.extraListeners=${spark2ExtraListeners}
                 --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                 --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                 --conf spark.sql.shuffle.partitions=3840
-                --conf spark.shuffle.sort.bypassMergeThreshold=3840
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
             <arg>--inputPath</arg><arg>${workingDir}/join_entities</arg>