1
0
Fork 0
This commit is contained in:
Miriam Baglioni 2021-06-18 09:45:16 +02:00
parent b486ae498f
commit 13c96622c9
1 changed files with 31 additions and 5 deletions

View File

@ -24,7 +24,7 @@
</parameters> </parameters>
<start to="generateCrossrefDataset"/> <start to="GenerateCrossrefDataset"/>
<kill name="Kill"> <kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message> <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
@ -40,11 +40,37 @@
<arg>--workingPath</arg><arg>${crossrefDumpPath}</arg> <arg>--workingPath</arg><arg>${crossrefDumpPath}</arg>
<arg>--outputPath</arg><arg>${workingDir}/files/</arg> <arg>--outputPath</arg><arg>${workingDir}/files/</arg>
</java> </java>
<ok to="generateCrossrefDataset"/> <ok to="UnpackCrossrefEntries"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="generateCrossrefDataset"> <action name="UnpackCrossrefEntries">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>SparkGenerateCrossrefDataset</name>
<class>eu.dnetlib.doiboost.crossref.UnpackCrtossrefEntries</class>
<jar>dhp-doiboost-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.sql.shuffle.partitions=3840
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--master</arg><arg>yarn-cluster</arg>
<arg>--sourcePath</arg><arg>${crossrefDumpPath}/files</arg>
<arg>--targetPath</arg><arg>${inputPathCrossref}/crossref_ds</arg>
</spark>
<ok to="GenerateCrossrefDataset"/>
<error to="Kill"/>
</action>
<action name="GenerateCrossrefDataset">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master> <master>yarn-cluster</master>
<mode>cluster</mode> <mode>cluster</mode>
@ -62,8 +88,8 @@
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts> </spark-opts>
<arg>--master</arg><arg>yarn-cluster</arg> <arg>--master</arg><arg>yarn-cluster</arg>
<arg>--sourcePath</arg><arg>${workingDir}/files</arg> <arg>--sourcePath</arg><arg>${inputPathCrossref}/crossref_ds</arg>
<arg>--targetPath</arg><arg>${inputPathCrossref}/crossref_ds</arg> <arg>--targetPath</arg><arg>${inputPathCrossref}/crossref_ds_updates</arg>
</spark> </spark>
<ok to="End"/> <ok to="End"/>