224 lines
9.1 KiB
XML
224 lines
9.1 KiB
XML
<workflow-app name="create RAW Organizations" xmlns="uri:oozie:workflow:0.5">
|
|
|
|
<parameters>
|
|
<property>
|
|
<name>graphOutputPath</name>
|
|
<description>the target path to store raw graph</description>
|
|
</property>
|
|
<property>
|
|
<name>reuseContent</name>
|
|
<value>false</value>
|
|
<description>should import content from the aggregator or reuse a previous version</description>
|
|
</property>
|
|
<property>
|
|
<name>contentPath</name>
|
|
<description>path location to store (or reuse) content from the aggregator</description>
|
|
</property>
|
|
<property>
|
|
<name>postgresURL</name>
|
|
<description>the postgres URL to access to the database</description>
|
|
</property>
|
|
<property>
|
|
<name>postgresUser</name>
|
|
<description>the user postgres</description>
|
|
</property>
|
|
<property>
|
|
<name>postgresPassword</name>
|
|
<description>the password postgres</description>
|
|
</property>
|
|
<property>
|
|
<name>postgresOpenOrgsURL</name>
|
|
<description>the postgres URL to access to the OpenOrgs database</description>
|
|
</property>
|
|
<property>
|
|
<name>postgresOpenOrgsUser</name>
|
|
<description>the user of OpenOrgs database</description>
|
|
</property>
|
|
<property>
|
|
<name>postgresOpenOrgsPassword</name>
|
|
<description>the password of OpenOrgs database</description>
|
|
</property>
|
|
<property>
|
|
<name>dbSchema</name>
|
|
<value>beta</value>
|
|
<description>the database schema according to the D-Net infrastructure (beta or production)</description>
|
|
</property>
|
|
<property>
|
|
<name>openOrgsAction</name>
|
|
<value>openorgs_dedup</value>
|
|
<description>The action to be executed during the import for OpenOrgs</description>
|
|
</property>
|
|
<property>
|
|
<name>isLookupUrl</name>
|
|
<description>the address of the lookUp service</description>
|
|
</property>
|
|
<property>
|
|
<name>nsPrefixBlacklist</name>
|
|
<value>foo</value>
|
|
<description>a blacklist of nsprefixes (comma separeted)</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>${oozieActionShareLibForSpark2}</value>
|
|
</property>
|
|
</configuration>
|
|
</global>
|
|
|
|
<start to="reuse_aggregator_content"/>
|
|
|
|
<kill name="Kill">
|
|
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
|
|
</kill>
|
|
|
|
<decision name="reuse_aggregator_content">
|
|
<switch>
|
|
<case to="ImportDB">${wf:conf('reuseContent') eq false}</case>
|
|
<case to="GenerateEntities">${wf:conf('reuseContent') eq true}</case>
|
|
<default to="ImportDB"/>
|
|
</switch>
|
|
</decision>
|
|
|
|
<fork name="ImportDB">
|
|
<path start="ImportDB_openaire_organizations"/>
|
|
<path start="ImportDB_openorgs"/>
|
|
</fork>
|
|
|
|
<action name="ImportDB_openaire_organizations">
|
|
<java>
|
|
<prepare>
|
|
<delete path="${contentPath}/db_openaire_organizations"/>
|
|
</prepare>
|
|
<main-class>eu.dnetlib.dhp.oa.graph.raw.MigrateDbEntitiesApplication</main-class>
|
|
<arg>--hdfsPath</arg><arg>${contentPath}/db_openaire_organizations</arg>
|
|
<arg>--postgresUrl</arg><arg>${postgresURL}</arg>
|
|
<arg>--postgresUser</arg><arg>${postgresUser}</arg>
|
|
<arg>--postgresPassword</arg><arg>${postgresPassword}</arg>
|
|
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
|
|
<arg>--action</arg><arg>openaire_organizations</arg>
|
|
<arg>--dbschema</arg><arg>${dbSchema}</arg>
|
|
<arg>--nsPrefixBlacklist</arg><arg>${nsPrefixBlacklist}</arg>
|
|
</java>
|
|
<ok to="wait_import"/>
|
|
<error to="Kill"/>
|
|
</action>
|
|
|
|
<action name="ImportDB_openorgs">
|
|
<java>
|
|
<prepare>
|
|
<delete path="${contentPath}/db_openorgs"/>
|
|
</prepare>
|
|
<main-class>eu.dnetlib.dhp.oa.graph.raw.MigrateDbEntitiesApplication</main-class>
|
|
<arg>--hdfsPath</arg><arg>${contentPath}/db_openorgs</arg>
|
|
<arg>--postgresUrl</arg><arg>${postgresOpenOrgsURL}</arg>
|
|
<arg>--postgresUser</arg><arg>${postgresOpenOrgsUser}</arg>
|
|
<arg>--postgresPassword</arg><arg>${postgresOpenOrgsPassword}</arg>
|
|
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
|
|
<arg>--action</arg><arg>${openOrgsAction}</arg>
|
|
<arg>--dbschema</arg><arg>${dbSchema}</arg>
|
|
<arg>--nsPrefixBlacklist</arg><arg>${nsPrefixBlacklist}</arg>
|
|
</java>
|
|
<ok to="wait_import"/>
|
|
<error to="Kill"/>
|
|
</action>
|
|
|
|
<join name="wait_import" to="GenerateEntities"/>
|
|
|
|
<action name="GenerateEntities">
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
<master>yarn</master>
|
|
<mode>cluster</mode>
|
|
<name>GenerateEntities</name>
|
|
<class>eu.dnetlib.dhp.oa.graph.raw.GenerateEntitiesApplication</class>
|
|
<jar>dhp-graph-mapper-${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}${spark2EventLogDir}
|
|
</spark-opts>
|
|
<arg>--sourcePaths</arg><arg>${contentPath}/db_openaire_organizations,${contentPath}/db_openorgs</arg>
|
|
<arg>--targetPath</arg><arg>${workingDir}/entities</arg>
|
|
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
|
|
</spark>
|
|
<ok to="GenerateGraph"/>
|
|
<error to="Kill"/>
|
|
</action>
|
|
|
|
<action name="GenerateGraph">
|
|
<spark xmlns="uri:oozie:spark-action:0.2">
|
|
<master>yarn</master>
|
|
<mode>cluster</mode>
|
|
<name>GenerateGraph</name>
|
|
<class>eu.dnetlib.dhp.oa.graph.raw.DispatchEntitiesApplication</class>
|
|
<jar>dhp-graph-mapper-${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}${spark2EventLogDir}
|
|
--conf spark.sql.shuffle.partitions=7680
|
|
</spark-opts>
|
|
<arg>--sourcePath</arg><arg>${workingDir}/entities</arg>
|
|
<arg>--graphRawPath</arg><arg>${graphOutputPath}</arg>
|
|
</spark>
|
|
<ok to="End"/>
|
|
<error to="Kill"/>
|
|
</action>
|
|
|
|
<end name="End"/>
|
|
|
|
</workflow-app> |