dnet-hadoop/dhp-workflows/dhp-aggregation/src/main/resources/eu/dnetlib/dhp/collection/oozie_app/workflow.xml

212 lines
8.6 KiB
XML

<workflow-app name="CollectionWorkflow" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>apiDescription</name>
<description>A json encoding of the API Description class</description>
</property>
<property>
<name>dataSourceInfo</name>
<description>A json encoding of the Datasource Info</description>
</property>
<property>
<name>identifierPath</name>
<description>An xpath to retrieve the metadata identifier for the generation of DNet Identifier </description>
</property>
<property>
<name>metadataEncoding</name>
<description> The type of the metadata XML/JSON</description>
</property>
<property>
<name>timestamp</name>
<description>The timestamp of the collection date</description>
</property>
<property>
<name>workflowId</name>
<description>The identifier of the workflow</description>
</property>
<property>
<name>mdStoreID</name>
<description>The identifier of the mdStore</description>
</property>
<property>
<name>mdStoreManagerURI</name>
<description>The URI of the MDStore Manager</description>
</property>
<property>
<name>dnetMessageManagerURL</name>
<description>The URI of the Dnet Message Manager</description>
</property>
<property>
<name>collectionMode</name>
<description>Should be REFRESH or INCREMENTAL</description>
</property>
<property>
<name>collection_java_xmx</name>
<value>-Xmx200m</value>
<description>Used to configure the heap size for the map JVM process. Should be 80% of mapreduce.map.memory.mb.</description>
</property>
</parameters>
<global>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
</global>
<start to="collection_mode"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<decision name="collection_mode">
<switch>
<case to="StartTransaction">${wf:conf('collectionMode') eq 'REFRESH'}</case>
<case to="BeginRead">${wf:conf('collectionMode') eq 'INCREMENTAL'}</case>
<default to="StartTransaction"/>
</switch>
</decision>
<action name="BeginRead">
<java>
<main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--action</arg><arg>READ_LOCK</arg>
<arg>--mdStoreID</arg><arg>${mdStoreID}</arg>
<arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
<capture-output/>
</java>
<ok to="StartTransaction"/>
<error to="Kill"/>
</action>
<action name="StartTransaction">
<java>
<main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--action</arg><arg>NEW_VERSION</arg>
<arg>--mdStoreID</arg><arg>${mdStoreID}</arg>
<arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
<capture-output/>
</java>
<ok to="CollectionWorker"/>
<error to="Kill"/>
</action>
<action name="CollectionWorker">
<java>
<main-class>eu.dnetlib.dhp.collection.CollectorWorkerApplication</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--apidescriptor</arg><arg>${apiDescription}</arg>
<arg>--namenode</arg><arg>${nameNode}</arg>
<arg>--workflowId</arg><arg>${workflowId}</arg>
<arg>--dnetMessageManagerURL</arg><arg>${dnetMessageManagerURL}</arg>
<arg>--mdStoreVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
<arg>--maxNumberOfRetry</arg><arg>${maxNumberOfRetry}</arg>
<arg>--requestDelay</arg><arg>${requestDelay}</arg>
<arg>--retryDelay</arg><arg>${retryDelay}</arg>
<arg>--connectTimeOut</arg><arg>${connectTimeOut}</arg>
<arg>--readTimeOut</arg><arg>${readTimeOut}</arg>
</java>
<ok to="GenerateNativeStoreSparkJob"/>
<error to="FailCollection"/>
</action>
<action name="GenerateNativeStoreSparkJob">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Generate Native MetadataStore</name>
<class>eu.dnetlib.dhp.collection.GenerateNativeStoreSparkJob</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${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}${spark2EventLogDir}
</spark-opts>
<arg>--encoding</arg><arg>${metadataEncoding}</arg>
<arg>--dateOfCollection</arg><arg>${timestamp}</arg>
<arg>--provenance</arg><arg>${dataSourceInfo}</arg>
<arg>--xpath</arg><arg>${identifierPath}</arg>
<arg>--mdStoreVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
<arg>--readMdStoreVersion</arg><arg>${wf:actionData('BeginRead')['mdStoreReadLockVersion']}</arg>
</spark>
<ok to="collection_mode_end"/>
<error to="FailCollection"/>
</action>
<decision name="collection_mode_end">
<switch>
<case to="CommitVersion">${wf:conf('collectionMode') eq 'REFRESH'}</case>
<case to="EndRead">${wf:conf('collectionMode') eq 'INCREMENTAL'}</case>
<default to="CommitVersion"/>
</switch>
</decision>
<action name="EndRead">
<java>
<main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--action</arg><arg>READ_UNLOCK</arg>
<arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
<arg>--readMDStoreId</arg><arg>${wf:actionData('BeginRead')['mdStoreReadLockVersion']}</arg>
</java>
<ok to="CommitVersion"/>
<error to="Kill"/>
</action>
<action name="CommitVersion">
<java>
<main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--action</arg><arg>COMMIT</arg>
<arg>--namenode</arg><arg>${nameNode}</arg>
<arg>--mdStoreVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
<arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
</java>
<ok to="End"/>
<error to="Kill"/>
</action>
<decision name="FailCollection">
<switch>
<case to="RollBack">${wf:conf('collectionMode') eq 'REFRESH'}</case>
<case to="EndReadRollBack">${wf:conf('collectionMode') eq 'INCREMENTAL'}</case>
<default to="RollBack"/>
</switch>
</decision>
<action name="EndReadRollBack">
<java>
<main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--action</arg><arg>READ_UNLOCK</arg>
<arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
<arg>--readMDStoreId</arg><arg>${wf:actionData('BeginRead')['mdStoreReadLockVersion']}</arg>
</java>
<ok to="RollBack"/>
<error to="Kill"/>
</action>
<action name="RollBack">
<java>
<main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
<java-opts>${collection_java_xmx}</java-opts>
<arg>--action</arg><arg>ROLLBACK</arg>
<arg>--mdStoreVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
<arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
</java>
<ok to="Kill"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>