2019-12-17 16:24:49 +01:00
|
|
|
<workflow-app name="index_infospace_graph" xmlns="uri:oozie:workflow:0.5">
|
2020-03-18 15:43:54 +01:00
|
|
|
|
2019-12-17 16:24:49 +01:00
|
|
|
<parameters>
|
2020-04-04 14:03:43 +02:00
|
|
|
<property>
|
|
|
|
<name>inputGraphRootPath</name>
|
|
|
|
<description>root location of input materialized graph</description>
|
|
|
|
</property>
|
2020-04-06 08:59:58 +02:00
|
|
|
<property>
|
|
|
|
<name>isLookupUrl</name>
|
|
|
|
<description>URL for the isLookup service</description>
|
|
|
|
</property>
|
2020-05-20 14:55:38 +02:00
|
|
|
<property>
|
|
|
|
<name>relPartitions</name>
|
|
|
|
<description>number or partitions for the relations Dataset</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>relationFilter</name>
|
|
|
|
<description>filter applied reading relations (by relClass)</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>maxRelations</name>
|
|
|
|
<description>maximum number of relations allowed for a each entity</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>otherDsTypeId</name>
|
|
|
|
<description>mapping used to populate datasourceTypeUi field</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>format</name>
|
|
|
|
<description>metadata format name (DMF|TMF)</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>batchSize</name>
|
|
|
|
<description>number of records to be included in each indexing request</description>
|
|
|
|
</property>
|
2020-04-04 14:03:43 +02:00
|
|
|
|
2019-12-17 16:24:49 +01:00
|
|
|
<property>
|
2020-04-01 19:07:30 +02:00
|
|
|
<name>sparkDriverMemoryForJoining</name>
|
|
|
|
<description>memory for driver process</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkExecutorMemoryForJoining</name>
|
|
|
|
<description>memory for individual executor</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
|
|
|
<name>sparkExecutorCoresForJoining</name>
|
|
|
|
<description>number of cores used by single executor</description>
|
2019-12-17 16:24:49 +01:00
|
|
|
</property>
|
|
|
|
<property>
|
2020-04-01 19:07:30 +02:00
|
|
|
<name>sparkDriverMemoryForIndexing</name>
|
2019-12-17 16:24:49 +01:00
|
|
|
<description>memory for driver process</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
2020-04-01 19:07:30 +02:00
|
|
|
<name>sparkExecutorMemoryForIndexing</name>
|
2019-12-17 16:24:49 +01:00
|
|
|
<description>memory for individual executor</description>
|
|
|
|
</property>
|
|
|
|
<property>
|
2020-04-01 19:07:30 +02:00
|
|
|
<name>sparkExecutorCoresForIndexing</name>
|
2019-12-17 16:24:49 +01:00
|
|
|
<description>number of cores used by single executor</description>
|
|
|
|
</property>
|
2020-04-02 18:44:09 +02:00
|
|
|
<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>
|
2020-01-21 16:35:44 +01:00
|
|
|
<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>
|
2020-05-21 10:00:14 +02:00
|
|
|
<property>
|
|
|
|
<name>sparkNetworkTimeout</name>
|
|
|
|
<description>configures spark.network.timeout</description>
|
|
|
|
</property>
|
2019-12-17 16:24:49 +01:00
|
|
|
</parameters>
|
|
|
|
|
2020-03-18 15:43:54 +01:00
|
|
|
<global>
|
|
|
|
<job-tracker>${jobTracker}</job-tracker>
|
|
|
|
<name-node>${nameNode}</name-node>
|
|
|
|
<configuration>
|
|
|
|
<property>
|
2020-04-02 18:44:09 +02:00
|
|
|
<name>oozie.action.sharelib.for.spark</name>
|
|
|
|
<value>${oozieActionShareLibForSpark2}</value>
|
2020-03-18 15:43:54 +01:00
|
|
|
</property>
|
|
|
|
</configuration>
|
|
|
|
</global>
|
|
|
|
|
2020-06-24 18:34:42 +02:00
|
|
|
<start to="prepare_relations"/>
|
2019-12-17 16:24:49 +01:00
|
|
|
|
|
|
|
<kill name="Kill">
|
|
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
|
|
</kill>
|
|
|
|
|
2020-04-04 14:03:43 +02:00
|
|
|
<action name="prepare_relations">
|
|
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
|
|
<master>yarn</master>
|
|
|
|
<mode>cluster</mode>
|
|
|
|
<name>PrepareRelations</name>
|
|
|
|
<class>eu.dnetlib.dhp.oa.provision.PrepareRelationsJob</class>
|
|
|
|
<jar>dhp-graph-provision-${projectVersion}.jar</jar>
|
|
|
|
<spark-opts>
|
|
|
|
--executor-cores=${sparkExecutorCoresForJoining}
|
|
|
|
--executor-memory=${sparkExecutorMemoryForJoining}
|
|
|
|
--driver-memory=${sparkDriverMemoryForJoining}
|
|
|
|
--conf spark.extraListeners=${spark2ExtraListeners}
|
|
|
|
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
|
|
|
|
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
|
|
|
|
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
|
2020-04-21 16:15:04 +02:00
|
|
|
--conf spark.sql.shuffle.partitions=3840
|
2020-04-04 14:03:43 +02:00
|
|
|
</spark-opts>
|
|
|
|
<arg>--inputRelationsPath</arg><arg>${inputGraphRootPath}/relation</arg>
|
|
|
|
<arg>--outputPath</arg><arg>${workingDir}/relation</arg>
|
2020-05-29 10:58:15 +02:00
|
|
|
<arg>--relPartitions</arg><arg>5000</arg>
|
2020-04-04 14:03:43 +02:00
|
|
|
</spark>
|
2020-06-24 18:34:42 +02:00
|
|
|
<ok to="End"/>
|
2020-04-04 14:03:43 +02:00
|
|
|
<error to="Kill"/>
|
|
|
|
</action>
|
|
|
|
|
|
|
|
|
2019-12-17 16:24:49 +01:00
|
|
|
|
|
|
|
<end name="End"/>
|
2020-05-20 14:55:38 +02:00
|
|
|
|
2019-12-17 16:24:49 +01:00
|
|
|
</workflow-app>
|