From 50c18f7a0b05940a476ed2ef900e15c329b7a398 Mon Sep 17 00:00:00 2001 From: Claudio Atzori Date: Tue, 30 Apr 2024 12:34:16 +0200 Subject: [PATCH] [dedup wf] revised memory settings to address the increased volume of input contents --- .../dedup/consistency/oozie_app/workflow.xml | 2 + .../dhp/oa/dedup/scan/oozie_app/workflow.xml | 46 ++++++------------- 2 files changed, 16 insertions(+), 32 deletions(-) diff --git a/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/consistency/oozie_app/workflow.xml b/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/consistency/oozie_app/workflow.xml index 306229e79..46dc71c2c 100644 --- a/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/consistency/oozie_app/workflow.xml +++ b/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/consistency/oozie_app/workflow.xml @@ -102,6 +102,8 @@ --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.sql.shuffle.partitions=15000 + --conf spark.network.timeout=300s + --conf spark.shuffle.registration.timeout=50000 --graphBasePath${graphBasePath} --graphOutputPath${graphOutputPath} diff --git a/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml b/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml index 49a331def..ff37c5074 100644 --- a/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml +++ b/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml @@ -33,16 +33,14 @@ max number of elements in a connected component - sparkDriverMemory - memory for driver process + sparkResourceOpts + --executor-memory=6G --conf spark.executor.memoryOverhead=4G --executor-cores=6 --driver-memory=8G --driver-cores=4 + spark resource options - sparkExecutorMemory - memory for individual executor - - - sparkExecutorCores - number of cores used by single executor + sparkResourceOptsCreateMergeRel + --executor-memory=6G --conf spark.executor.memoryOverhead=4G --executor-cores=6 --driver-memory=8G --driver-cores=4 + spark resource options oozieActionShareLibForSpark2 @@ -119,9 +117,7 @@ eu.dnetlib.dhp.oa.dedup.SparkCreateSimRels dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -146,9 +142,7 @@ eu.dnetlib.dhp.oa.dedup.SparkWhitelistSimRels dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -174,9 +168,7 @@ eu.dnetlib.dhp.oa.dedup.SparkCreateMergeRels dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOptsCreateMergeRel} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -203,9 +195,7 @@ eu.dnetlib.dhp.oa.dedup.SparkCreateDedupRecord dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -230,9 +220,7 @@ eu.dnetlib.dhp.oa.dedup.SparkCopyOpenorgsMergeRels dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -257,9 +245,7 @@ eu.dnetlib.dhp.oa.dedup.SparkCreateOrgsDedupRecord dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -283,9 +269,7 @@ eu.dnetlib.dhp.oa.dedup.SparkUpdateEntity dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} @@ -309,9 +293,7 @@ eu.dnetlib.dhp.oa.dedup.SparkCopyRelationsNoOpenorgs dhp-dedup-openaire-${projectVersion}.jar - --executor-memory=${sparkExecutorMemory} - --executor-cores=${sparkExecutorCores} - --driver-memory=${sparkDriverMemory} + ${sparkResourceOpts} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}