added hive parameter and a step of reset of the working dir in the workflow

This commit is contained in:
Miriam Baglioni 2020-04-16 18:42:04 +02:00
parent 6afbd542ca
commit a9120164aa
2 changed files with 24 additions and 45 deletions

View File

@ -19,6 +19,10 @@
<name>spark2YarnHistoryServerAddress</name>
<value>http://iis-cdh5-test-gw.ocean.icm.edu.pl:18089</value>
</property>
<property>
<name>hive_metastore_uris</name>
<value>thrift://iis-cdh5-test-m3.ocean.icm.edu.pl:9083</value>
</property>
<property>
<name>spark2EventLogDir</name>
<value>/user/spark/spark2ApplicationHistory</value>

View File

@ -17,13 +17,21 @@
<description>writes new version of the graph after the propagation step</description>
</property>
</parameters>
<start to="reset-outputpath"/>
<start to="fork_prepare_assoc_step1"/>
<!-- <start to="fork_prepare_assoc_step1"/>-->
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="reset-outputpath">
<fs>
<delete path='${workingDir}/projecttoresult_propagation/relation'/>
</fs>
<ok to="fork_prepare_assoc_step1"/>
<error to="Kill"/>
</action>
<fork name="fork_prepare_assoc_step1">
<path start="join_prepare_publication"/>
<path start="join_prepare_dataset"/>
@ -35,7 +43,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-PreparePhase1-Publications</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.PrepareResultOrcidAssociationStep1</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -63,7 +71,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-PreparePhase1-Dataset</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.PrepareResultOrcidAssociationStep1</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -91,7 +99,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-PreparePhase1-ORP</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.PrepareResultOrcidAssociationStep1</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -119,7 +127,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-PreparePhase1-Software</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.PrepareResultOrcidAssociationStep1</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -139,40 +147,7 @@
<arg>--outputPath</arg><arg>${workingDir}/preparedInfo/targetOrcidAssoc</arg>
<arg>--allowedsemrels</arg><arg>${allowedsemrels}</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>
<action name="OrcidToResultFromSemRelPropagation">
<spark xmlns="uri:oozie:spark-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>OrcidToResultFromSemRelPropagation</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.SparkOrcidToResultFromSemRelJob</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
--num-executors=${sparkExecutorNumber}
--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.dynamicAllocation.enabled=true
--conf spark.dynamicAllocation.maxExecutors=${spark2MaxExecutors}
</spark-opts>
<arg>-mt</arg> <arg>yarn-cluster</arg>
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--allowedsemrels</arg><arg>${allowedsemrels}</arg>
<arg>--writeUpdate</arg><arg>${writeUpdate}</arg>
<arg>--saveGraph</arg><arg>${saveGraph}</arg>
</spark>
<ok to="End"/>
<ok to="wait"/>
<error to="Kill"/>
</action>
@ -182,7 +157,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-PreparePhase2</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.PrepareResultOrcidAssociationStep2</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -215,7 +190,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-Publication</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.SparkOrcidToResultFromSemRelJob3</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -241,7 +216,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-Dataset</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.SparkOrcidToResultFromSemRelJob3</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -267,7 +242,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-ORP</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.SparkOrcidToResultFromSemRelJob3</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>
@ -293,7 +268,7 @@
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>ORCIDpropagationFromSemRel</name>
<name>ORCIDPropagation-Software</name>
<class>eu.dnetlib.dhp.orcidtoresultfromsemrel.SparkOrcidToResultFromSemRelJob3</class>
<jar>dhp-propagation-${projectVersion}.jar</jar>
<spark-opts>