dnet-hadoop/dhp-workflows/dhp-graph-provision-scholex.../src/main/resources/eu/dnetlib/dhp/sx/synch/oozie_app/workflow.xml

97 lines
3.8 KiB
XML

<workflow-app name="Keep On Synch datacite" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>workingDirPath</name>
<description>the source path</description>
</property>
<property>
<name>sparkDriverMemory</name>
<description>memory for driver process</description>
</property>
<property>
<name>sparkExecutorMemory</name>
<description>memory for individual executor</description>
</property>
<property>
<name>index</name>
<description>index name</description>
</property>
<property>
<name>timestamp</name>
<description>timestamp from incremental harvesting</description>
</property>
</parameters>
<start to="ResetWorkingPath"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="ResetWorkingPath">
<fs>
<delete path='${workingDirPath}/synch'/>
<mkdir path='${workingDirPath}/synch'/>
</fs>
<ok to="ImportDataciteUpdate"/>
<error to="Kill"/>
</action>
<action name="ImportDataciteUpdate">
<java>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<main-class>eu.dnetlib.dhp.provision.update.RetrieveUpdateFromDatacite</main-class>
<arg>-t</arg><arg>${workingDirPath}/synch/input_json</arg>
<arg>-n</arg><arg>${nameNode}</arg>
<arg>-ts</arg><arg>${timestamp}</arg>
<arg>-ih</arg><arg>ip-90-147-167-25.ct1.garrservices.it</arg>
<arg>-in</arg><arg>datacite</arg>
</java>
<ok to="resolveScholix"/>
<error to="Kill"/>
</action>
<action name="resolveScholix">
<spark xmlns="uri:oozie:spark-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>resolve and generate Scholix</name>
<class>eu.dnetlib.dhp.provision.update.SparkResolveScholixTarget</class>
<jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>
<spark-opts>--executor-memory ${sparkExecutorMemory} --driver-memory=${sparkDriverMemory} ${sparkExtraOPT} --conf spark.dynamicAllocation.maxExecutors="32" </spark-opts>
<arg>-m</arg> <arg>yarn-cluster</arg>
<arg>-s</arg><arg>${workingDirPath}/synch/input_json</arg>
<arg>-w</arg><arg>${workingDirPath}/synch</arg>
<arg>-h</arg><arg>ip-90-147-167-25.ct1.garrservices.it</arg>
</spark>
<ok to="indexScholix"/>
<error to="Kill"/>
</action>
<action name="indexScholix">
<spark xmlns="uri:oozie:spark-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>index scholix</name>
<class>eu.dnetlib.dhp.provision.SparkIndexCollectionOnES</class>
<jar>dhp-graph-provision-scholexplorer-${projectVersion}.jar</jar>
<spark-opts>--executor-memory ${sparkExecutorMemory} --driver-memory=${sparkDriverMemory} ${sparkExtraOPT} --conf spark.dynamicAllocation.maxExecutors="8" </spark-opts>
<arg>-mt</arg> <arg>yarn-cluster</arg>
<arg>--sourcePath</arg><arg>${workingDirPath}/synch/resolved_json</arg>
<arg>--index</arg><arg>${index}_scholix</arg>
<arg>--idPath</arg><arg>identifier</arg>
<arg>--type</arg><arg>scholix</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>