dnet-hadoop/dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/consistency/oozie_app/workflow.xml

<workflow-app name="Update Graph" xmlns="uri:oozie:workflow:0.5">
    <parameters>
        <property>
            <name>graphBasePath</name>
            <description>the raw graph base path</description>
        </property>
        <property>
            <name>workingPath</name>
            <description>path of the working directory</description>
        </property>
        <property>
            <name>dedupGraphPath</name>
            <description>path of the dedup graph</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>sparkExecutorCores</name>
            <description>number of cores used by single executor</description>
        </property>

        <property>
            <name>oozieActionShareLibForSpark2</name>
            <description>oozie action sharelib for spark 2.*</description>
        </property>
        <property>
            <name>spark2ExtraListeners</name>
            <value>com.cloudera.spark.lineage.NavigatorAppListener</value>
            <description>spark 2.* extra listeners classname</description>
        </property>
        <property>
            <name>spark2SqlQueryExecutionListeners</name>
            <value>com.cloudera.spark.lineage.NavigatorQueryListener</value>
            <description>spark 2.* sql query execution listeners classname</description>
        </property>
        <property>
            <name>spark2YarnHistoryServerAddress</name>
            <description>spark 2.* yarn history server address</description>
        </property>
        <property>
            <name>spark2EventLogDir</name>
            <description>spark 2.* event log dir location</description>
        </property>
    </parameters>

    <global>
        <job-tracker>${jobTracker}</job-tracker>
        <name-node>${nameNode}</name-node>
        <configuration>
            <property>
                <name>mapreduce.job.queuename</name>
                <value>${queueName}</value>
            </property>
            <property>
                <name>oozie.launcher.mapred.job.queue.name</name>
                <value>${oozieLauncherQueueName}</value>
            </property>
            <property>
                <name>oozie.action.sharelib.for.spark</name>
                <value>${oozieActionShareLibForSpark2}</value>
            </property>
        </configuration>
    </global>

    <start to="PropagateRelation"/>

    <kill name="Kill">
        <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
    </kill>

    <action name="PropagateRelation">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>Update Relations</name>
            <class>eu.dnetlib.dhp.oa.dedup.SparkPropagateRelation</class>
            <jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${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}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=7680
            </spark-opts>
            <arg>--i</arg><arg>${graphBasePath}</arg>
            <arg>--o</arg><arg>${dedupGraphPath}</arg>
            <arg>--w</arg><arg>${workingPath}</arg>
        </spark>
        <ok to="fork_copy_entities"/>
        <error to="Kill"/>
    </action>

    <fork name="fork_copy_entities">
        <path start="copy_datasource"/>
        <path start="copy_project"/>
        <path start="copy_organization"/>
        <path start="copy_publication"/>
        <path start="copy_dataset"/>
        <path start="copy_software"/>
        <path start="copy_otherresearchproduct"/>
    </fork>

    <action name="copy_datasource">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/datasource"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/datasource</arg>
            <arg>${dedupGraphPath}/datasource</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <action name="copy_project">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/project"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/project</arg>
            <arg>${dedupGraphPath}/project</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <action name="copy_organization">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/organization"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/organization</arg>
            <arg>${dedupGraphPath}/organization</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <action name="copy_publication">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/publication"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/publication</arg>
            <arg>${dedupGraphPath}/publication</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <action name="copy_dataset">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/dataset"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/dataset</arg>
            <arg>${dedupGraphPath}/dataset</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <action name="copy_software">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/software"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/software</arg>
            <arg>${dedupGraphPath}/software</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <action name="copy_otherresearchproduct">
        <distcp xmlns="uri:oozie:distcp-action:0.2">
            <prepare>
                <delete path="${dedupGraphPath}/otherresearchproduct"/>
            </prepare>
            <arg>-pb</arg>
            <arg>${graphBasePath}/otherresearchproduct</arg>
            <arg>${dedupGraphPath}/otherresearchproduct</arg>
        </distcp>
        <ok to="wait_copy"/>
        <error to="Kill"/>
    </action>

    <join name="wait_copy" to="End"/>

    <end name="End"/>
</workflow-app>