dnet-hadoop/dhp-workflows/dhp-broker-events/src/main/resources/eu/dnetlib/dhp/broker/oa/generate_all/oozie_app/workflow.xml

<workflow-app name="create broker events" xmlns="uri:oozie:workflow:0.5">

    <parameters>
        <property>
            <name>graphInputPath</name>
            <description>the path where the graph is stored</description>
        </property>
        <property>
            <name>outputDir</name>
            <description>the path where the the generated data will be stored</description>
        </property>
		<property>
            <name>datasourceIdWhitelist</name>
            <value>-</value>
            <description>a white list (comma separeted, - for empty list) of datasource ids</description>
        </property>
		<property>
            <name>datasourceTypeWhitelist</name>
            <value>-</value>
            <description>a white list (comma separeted, - for empty list) of datasource types</description>
        </property>
		<property>
            <name>datasourceIdBlacklist</name>
            <value>-</value>
            <description>a black list (comma separeted, - for empty list) of datasource ids</description>
        </property>
        <property>
            <name>topicWhitelist</name>
            <value>*</value>
            <description>a white list (comma separeted, * for all) of topics</description>
        </property>
        <property>
            <name>esEventIndexName</name>
            <description>the elasticsearch index name for events</description>
        </property>
        <property>
            <name>esNotificationsIndexName</name>
            <description>the elasticsearch index name for notifications</description>
        </property>
        <property>
            <name>esIndexHost</name>
            <description>the elasticsearch host</description>
        </property>
        <property>
            <name>esBatchWriteRetryCount</name>
            <value>8</value>
            <description>an ES configuration property</description>
        </property>
		<property>
            <name>esBatchWriteRetryWait</name>
            <value>60s</value>
            <description>an ES configuration property</description>
        </property>
		<property>
            <name>esBatchSizeEntries</name>
            <value>200</value>
            <description>an ES configuration property</description>
        </property>
		<property>
            <name>esNodesWanOnly</name>
            <value>true</value>
            <description>an ES configuration property</description>
        </property>
        <property>
        	<name>maxIndexedEventsForDsAndTopic</name>
        	<description>the max number of events for each couple (ds/topic)</description>
        </property>
        <property>
        	<name>brokerApiBaseUrl</name>
        	<description>the url of the broker service api</description>
        </property>
        <property>
        	<name>brokerDbUrl</name>
        	<description>the url of the broker database</description>
        </property>
        <property>
        	<name>brokerDbUser</name>
        	<description>the user of the broker database</description>
        </property>
        <property>
        	<name>brokerDbPassword</name>
        	<description>the password of the broker database</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>
        <property>
        	<name>sparkMaxExecutorsForIndexing</name>
        	<value>8</value>
        	<description>Max number of workers for ElasticSearch indexing</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="resume_from"/>

    <decision name="resume_from">
        <switch>
        	<case to="ensure_output_dir">${wf:conf('resumeFrom') eq 'ensure_output_dir'}</case>
            <case to="index_event_subset">${wf:conf('resumeFrom') eq 'index_event_subset'}</case>
            <case to="stats">${wf:conf('resumeFrom') eq 'stats'}</case>
            <case to="index_notifications">${wf:conf('resumeFrom') eq 'index_notifications'}</case>
            <default to="ensure_output_dir"/>
        </switch>
    </decision>

    <kill name="Kill">
        <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
    </kill>
    
    <action name="ensure_output_dir">
        <fs>
            <mkdir path='${outputDir}'/>
        </fs>
        <ok to="start_entities_and_rels"/>
        <error to="Kill"/>
    </action>
    
    <fork name="start_entities_and_rels">
		<path start="prepare_simple_entities"/>
		<path start="prepare_related_datasources"/>
        <path start="prepare_related_softwares"/> 
        <path start="prepare_related_datasets"/>
        <path start="prepare_related_projects"/>
        <path start="prepare_related_publications"/>
    </fork>
    
    <action name="prepare_simple_entities">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareSimpleEntititiesJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareSimpleEntititiesJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="wait_entities_and_rels"/>
        <error to="Kill"/>
    </action>
    
    <action name="prepare_related_datasources">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareRelatedDatasourcesJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareRelatedDatasourcesJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="wait_entities_and_rels"/>
        <error to="Kill"/>
    </action>

 
    <action name="prepare_related_datasets">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareRelatedDatasetsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareRelatedDatasetsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="wait_entities_and_rels"/>
        <error to="Kill"/>
    </action>

    <action name="prepare_related_projects">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareRelatedProjectsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareRelatedProjectsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="wait_entities_and_rels"/>
        <error to="Kill"/>
    </action>

    <action name="prepare_related_publications">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareRelatedPublicationsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareRelatedPublicationsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="wait_entities_and_rels"/>
        <error to="Kill"/>
    </action>
  
    <action name="prepare_related_softwares">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareRelatedSoftwaresJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareRelatedSoftwaresJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="wait_entities_and_rels"/>
        <error to="Kill"/>
    </action>
 
	<join name="wait_entities_and_rels" to="join_entities_step0"/>
    
    <action name="join_entities_step0">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>JoinStep0</name>
            <class>eu.dnetlib.dhp.broker.oa.JoinStep0Job</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="join_entities_step1"/>
        <error to="Kill"/>
    </action>
    
    <action name="join_entities_step1">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>JoinStep1</name>
            <class>eu.dnetlib.dhp.broker.oa.JoinStep1Job</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="join_entities_step2"/>
        <error to="Kill"/>
    </action>
    
    <action name="join_entities_step2">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>JoinStep2</name>
            <class>eu.dnetlib.dhp.broker.oa.JoinStep2Job</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="join_entities_step3"/>
        <error to="Kill"/>
    </action>
    
    <action name="join_entities_step3">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>JoinStep3</name>
            <class>eu.dnetlib.dhp.broker.oa.JoinStep3Job</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="join_entities_step4"/>
        <error to="Kill"/>
    </action>
    
    <action name="join_entities_step4">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>JoinStep4</name>
            <class>eu.dnetlib.dhp.broker.oa.JoinStep4Job</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="prepare_groups"/>
        <error to="Kill"/>
    </action>

    <action name="prepare_groups">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>PrepareGroupsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.PrepareGroupsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--graphPath</arg><arg>${graphInputPath}</arg>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
        </spark>
        <ok to="generate_events"/>
        <error to="Kill"/>
    </action>

    <action name="generate_events">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>GenerateEventsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.GenerateEventsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--workingDir</arg><arg>${workingDir}</arg>
            <arg>--outputDir</arg><arg>${outputDir}</arg>
			<arg>--datasourceIdWhitelist</arg><arg>${datasourceIdWhitelist}</arg>
			<arg>--datasourceTypeWhitelist</arg><arg>${datasourceTypeWhitelist}</arg>
			<arg>--datasourceIdBlacklist</arg><arg>${datasourceIdBlacklist}</arg>
			<arg>--topicWhitelist</arg><arg>${topicWhitelist}</arg>
        </spark>
        <ok to="index_event_subset"/>
        <error to="Kill"/>
    </action>
    
     <action name="index_event_subset">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>IndexEventSubsetOnESJob</name>
            <class>eu.dnetlib.dhp.broker.oa.IndexEventSubsetJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.dynamicAllocation.maxExecutors=${sparkMaxExecutorsForIndexing} 
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--outputDir</arg><arg>${outputDir}</arg>
            <arg>--index</arg><arg>${esEventIndexName}</arg>
            <arg>--esHost</arg><arg>${esIndexHost}</arg>
            <arg>--esBatchWriteRetryCount</arg><arg>${esBatchWriteRetryCount}</arg>
            <arg>--esBatchWriteRetryWait</arg><arg>${esBatchWriteRetryWait}</arg>
            <arg>--esBatchSizeEntries</arg><arg>${esBatchSizeEntries}</arg>
            <arg>--esNodesWanOnly</arg><arg>${esNodesWanOnly}</arg>
            <arg>--maxEventsForTopic</arg><arg>${maxIndexedEventsForDsAndTopic}</arg>
            <arg>--brokerApiBaseUrl</arg><arg>${brokerApiBaseUrl}</arg>
        </spark>
        <ok to="stats"/>
        <error to="Kill"/>
    </action>
    
    <action name="stats">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>GenerateStatsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.GenerateStatsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--outputDir</arg><arg>${outputDir}</arg>
            <arg>--dbUrl</arg><arg>${brokerDbUrl}</arg>
            <arg>--dbUser</arg><arg>${brokerDbUser}</arg>
            <arg>--dbPassword</arg><arg>${brokerDbPassword}</arg>
            <arg>--brokerApiBaseUrl</arg><arg>${brokerApiBaseUrl}</arg>
        </spark>
        <ok to="generate_notifications"/>
        <error to="Kill"/>
    </action>
    
    <action name="generate_notifications">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>GenerateNotificationsJob</name>
            <class>eu.dnetlib.dhp.broker.oa.GenerateNotificationsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-cores=${sparkExecutorCores}
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--outputDir</arg><arg>${outputDir}</arg>
            <arg>--brokerApiBaseUrl</arg><arg>${brokerApiBaseUrl}</arg>
        </spark>
        <ok to="index_notifications"/>
        <error to="Kill"/>
    </action>
    
    <action name="index_notifications">
        <spark xmlns="uri:oozie:spark-action:0.2">
            <master>yarn</master>
            <mode>cluster</mode>
            <name>IndexNotificationsOnESJob</name>
            <class>eu.dnetlib.dhp.broker.oa.IndexNotificationsJob</class>
            <jar>dhp-broker-events-${projectVersion}.jar</jar>
            <spark-opts>
                --executor-memory=${sparkExecutorMemory}
                --driver-memory=${sparkDriverMemory}
                --conf spark.dynamicAllocation.maxExecutors=${sparkMaxExecutorsForIndexing} 
                --conf spark.extraListeners=${spark2ExtraListeners}
                --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
                --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
                --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
                --conf spark.sql.shuffle.partitions=3840
            </spark-opts>
            <arg>--outputDir</arg><arg>${outputDir}</arg>
            <arg>--index</arg><arg>${esNotificationsIndexName}</arg>
            <arg>--esHost</arg><arg>${esIndexHost}</arg>
            <arg>--esBatchWriteRetryCount</arg><arg>${esBatchWriteRetryCount}</arg>
            <arg>--esBatchWriteRetryWait</arg><arg>${esBatchWriteRetryWait}</arg>
            <arg>--esBatchSizeEntries</arg><arg>${esBatchSizeEntries}</arg>
            <arg>--esNodesWanOnly</arg><arg>${esNodesWanOnly}</arg>
            <arg>--brokerApiBaseUrl</arg><arg>${brokerApiBaseUrl}</arg>
        </spark>
        <ok to="End"/>
        <error to="Kill"/>
    </action>
    
    <end name="End"/>

</workflow-app>