dnet-hadoop/dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/main/oozie_app/workflow.xml

275 lines
11 KiB
XML

<workflow-app name="promote_actionsets_for_graph" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>activePromoteDatasetActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Dataset payload</description>
</property>
<property>
<name>activePromoteDatasourceActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Datasource payload</description>
</property>
<property>
<name>activePromoteOrganizationActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Organization payload</description>
</property>
<property>
<name>activePromoteOtherResearchProductActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.OtherResearchProduct payload</description>
</property>
<property>
<name>activePromoteProjectActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Project payload</description>
</property>
<property>
<name>activePromotePublicationActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Publication payload</description>
</property>
<property>
<name>activePromoteRelationActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Relation payload</description>
</property>
<property>
<name>activePromoteResultActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Result payload</description>
</property>
<property>
<name>activePromoteSoftwareActionPayload</name>
<description>when true will promote actions with eu.dnetlib.dhp.schema.oaf.Software payload</description>
</property>
<property>
<name>inputGraphRootPath</name>
<description>root location of input materialized graph</description>
</property>
<property>
<name>isLookupUrl</name>
<description>URL of the ISLookupService</description>
</property>
<property>
<name>inputActionSetIds</name>
<description>comma separated list of action set ids to promote</description>
</property>
<property>
<name>outputGraphRootPath</name>
<description>root location for output materialized graph</description>
</property>
<property>
<name>mergeAndGetStrategy</name>
<description>strategy for merging graph table objects with action payload instances, MERGE_FROM_AND_GET or SELECT_NEWER_AND_GET</description>
</property>
<property>
<name>shouldGroupById</name>
<value>false</value>
<description>indicates whether the promotion operation should group objects in the graph by id or not</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>oozie.action.sharelib.for.spark</name>
<value>${oozieActionShareLibForSpark2}</value>
</property>
</configuration>
</global>
<start to="PartitionActionSetsByPayloadType"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="PartitionActionSetsByPayloadType">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>PartitionActionSetsByPayloadType</name>
<class>eu.dnetlib.dhp.actionmanager.partition.PartitionActionSetsByPayloadTypeJob</class>
<jar>dhp-actionmanager-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${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}${spark2EventLogDir}
</spark-opts>
<arg>--inputActionSetIds</arg><arg>${inputActionSetIds}</arg>
<arg>--outputPath</arg><arg>${workingDir}/action_payload_by_type</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
</spark>
<ok to="ForkPromote"/>
<error to="Kill"/>
</action>
<fork name="ForkPromote">
<path start="PromoteActionPayloadForDatasetTable"/>
<path start="PromoteActionPayloadForDatasourceTable"/>
<path start="PromoteActionPayloadForOrganizationTable"/>
<path start="PromoteActionPayloadForOtherResearchProductTable"/>
<path start="PromoteActionPayloadForProjectTable"/>
<path start="PromoteActionPayloadForPublicationTable"/>
<path start="PromoteActionPayloadForRelationTable"/>
<path start="PromoteActionPayloadForSoftwareTable"/>
</fork>
<action name="PromoteActionPayloadForDatasetTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_dataset_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForDatasourceTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_datasource_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForOrganizationTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_organization_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForOtherResearchProductTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_otherresearchproduct_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForProjectTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_project_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForPublicationTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_publication_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForRelationTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_relation_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<action name="PromoteActionPayloadForSoftwareTable">
<sub-workflow>
<app-path>${wf:appPath()}/promote_action_payload_for_software_table</app-path>
<propagate-configuration/>
<configuration>
<property>
<name>inputActionPayloadRootPath</name>
<value>${workingDir}/action_payload_by_type</value>
</property>
</configuration>
</sub-workflow>
<ok to="JoinPromote"/>
<error to="Kill"/>
</action>
<join name="JoinPromote" to="End"/>
<end name="End"/>
</workflow-app>