graphOutputPath the target path to store raw graph reuseContent false should import content from the aggregator or reuse a previous version contentPath path location to store (or reuse) content from the aggregator postgresURL the postgres URL to access to the database postgresUser the user postgres postgresPassword the password postgres postgresOpenOrgsURL the postgres URL to access to the OpenOrgs database postgresOpenOrgsUser the user of OpenOrgs database postgresOpenOrgsPassword the password of OpenOrgs database dbSchema beta the database schema according to the D-Net infrastructure (beta or production) isLookupUrl the address of the lookUp service nsPrefixBlacklist a blacklist of nsprefixes (comma separeted) sparkDriverMemory memory for driver process sparkExecutorMemory memory for individual executor sparkExecutorCores number of cores used by single executor oozieActionShareLibForSpark2 oozie action sharelib for spark 2.* spark2ExtraListeners com.cloudera.spark.lineage.NavigatorAppListener spark 2.* extra listeners classname spark2SqlQueryExecutionListeners com.cloudera.spark.lineage.NavigatorQueryListener spark 2.* sql query execution listeners classname spark2YarnHistoryServerAddress spark 2.* yarn history server address spark2EventLogDir spark 2.* event log dir location ${jobTracker} ${nameNode} mapreduce.job.queuename ${queueName} oozie.launcher.mapred.job.queue.name ${oozieLauncherQueueName} oozie.action.sharelib.for.spark ${oozieActionShareLibForSpark2} Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}] ${wf:conf('reuseContent') eq false} ${wf:conf('reuseContent') eq true} eu.dnetlib.dhp.oa.graph.raw.MigrateDbEntitiesApplication --hdfsPath${contentPath}/db_openaire_organizations --postgresUrl${postgresURL} --postgresUser${postgresUser} --postgresPassword${postgresPassword} --isLookupUrl${isLookupUrl} --actionopenaire_organizations --dbschema${dbSchema} --nsPrefixBlacklist${nsPrefixBlacklist} eu.dnetlib.dhp.oa.graph.raw.MigrateDbEntitiesApplication --hdfsPath${contentPath}/db_openorgs --postgresUrl${postgresOpenOrgsURL} --postgresUser${postgresOpenOrgsUser} --postgresPassword${postgresOpenOrgsPassword} --isLookupUrl${isLookupUrl} --actionopenorgs_dedup --dbschema${dbSchema} --nsPrefixBlacklist${nsPrefixBlacklist} yarn cluster GenerateEntities eu.dnetlib.dhp.oa.graph.raw.GenerateEntitiesApplication dhp-graph-mapper-${projectVersion}.jar --executor-memory ${sparkExecutorMemory} --executor-cores ${sparkExecutorCores} --driver-memory=${sparkDriverMemory} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --sourcePaths${contentPath}/db_openaire_organizations,${contentPath}/db_openorgs --targetPath${workingDir}/entities --isLookupUrl${isLookupUrl} yarn cluster GenerateGraph eu.dnetlib.dhp.oa.graph.raw.DispatchEntitiesApplication dhp-graph-mapper-${projectVersion}.jar --executor-memory ${sparkExecutorMemory} --executor-cores ${sparkExecutorCores} --driver-memory=${sparkDriverMemory} --conf spark.extraListeners=${spark2ExtraListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.sql.shuffle.partitions=7680 --sourcePath${workingDir}/entities --graphRawPath${graphOutputPath}