1
0
Fork 0

updated parameters for regular_all_steps worfklow

This commit is contained in:
Claudio Atzori 2020-03-17 15:23:54 +01:00
parent aeb01fa353
commit 4706f24ec5
1 changed files with 32 additions and 19 deletions

View File

@ -1,16 +1,13 @@
<workflow-app name="import regular entities as Graph (all steps)" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>migrationPathStep1</name>
<description>the base path to store hdfs file</description>
<name>workingPath</name>
<value>/tmp/dhp_migration</value>
<description>the base path to store temporary intermediate data</description>
</property>
<property>
<name>migrationPathStep2</name>
<description>the temporary path to store entities before dispatching</description>
</property>
<property>
<name>migrationPathStep3</name>
<description>the graph Raw base path</description>
<name>graphBasePath</name>
<description>the target path to store raw graph</description>
</property>
<property>
<name>postgresURL</name>
@ -54,8 +51,10 @@
<action name="ResetWorkingPath">
<fs>
<delete path='${migrationPathStep1}'/>
<mkdir path='${migrationPathStep1}'/>
<delete path='${workingPath}'/>
<mkdir path='${workingPath}'/>
<delete path="'${graphBasePath}/graph_raw'"/>
<mkdir path="'${graphBasePath}/graph_raw'"/>
</fs>
<ok to="ImportDB"/>
<error to="Kill"/>
@ -66,7 +65,7 @@
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<main-class>eu.dnetlib.dhp.migration.step1.MigrateDbEntitiesApplication</main-class>
<arg>-p</arg><arg>${migrationPathStep1}/db_records</arg>
<arg>-p</arg><arg>${workingPath}/db_records</arg>
<arg>-pgurl</arg><arg>${postgresURL}</arg>
<arg>-pguser</arg><arg>${postgresUser}</arg>
<arg>-pgpasswd</arg><arg>${postgresPassword}</arg>
@ -80,7 +79,7 @@
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<main-class>eu.dnetlib.dhp.migration.step1.MigrateMongoMdstoresApplication</main-class>
<arg>-p</arg><arg>${migrationPathStep1}/odf_records</arg>
<arg>-p</arg><arg>${workingPath}/odf_records</arg>
<arg>-mongourl</arg><arg>${mongoURL}</arg>
<arg>-mongodb</arg><arg>${mongoDb}</arg>
<arg>-f</arg><arg>ODF</arg>
@ -96,7 +95,7 @@
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<main-class>eu.dnetlib.dhp.migration.step1.MigrateMongoMdstoresApplication</main-class>
<arg>-p</arg><arg>${migrationPathStep1}/oaf_records</arg>
<arg>-p</arg><arg>${workingPath}/oaf_records</arg>
<arg>-mongourl</arg><arg>${mongoURL}</arg>
<arg>-mongodb</arg><arg>${mongoDb}</arg>
<arg>-f</arg><arg>OAF</arg>
@ -125,10 +124,17 @@
<name>GenerateEntities</name>
<class>eu.dnetlib.dhp.migration.step2.GenerateEntitiesApplication</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>--executor-memory ${sparkExecutorMemory} --executor-cores ${sparkExecutorCores} --driver-memory=${sparkDriverMemory} --conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener" --conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener" --conf spark.sql.warehouse.dir="/user/hive/warehouse"</spark-opts>
<spark-opts>
--executor-memory ${sparkExecutorMemory}
--executor-cores ${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
--conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
--conf spark.sql.warehouse.dir="/user/hive/warehouse"
</spark-opts>
<arg>-mt</arg> <arg>yarn-cluster</arg>
<arg>-s</arg><arg>${migrationPathStep1}/db_records,${migrationPathStep1}/oaf_records,${migrationPathStep1}/odf_records</arg>
<arg>-t</arg><arg>${migrationPathStep2}/all_entities</arg>
<arg>-s</arg><arg>${workingPath}/db_records,${workingPath}/oaf_records,${workingPath}/odf_records</arg>
<arg>-t</arg><arg>${workingPath}/all_entities</arg>
<arg>-pgurl</arg><arg>${postgresURL}</arg>
<arg>-pguser</arg><arg>${postgresUser}</arg>
<arg>-pgpasswd</arg><arg>${postgresPassword}</arg>
@ -155,10 +161,17 @@
<name>GenerateGraph</name>
<class>eu.dnetlib.dhp.migration.step3.DispatchEntitiesApplication</class>
<jar>dhp-aggregation-${projectVersion}.jar</jar>
<spark-opts>--executor-memory ${sparkExecutorMemory} --executor-cores ${sparkExecutorCores} --driver-memory=${sparkDriverMemory} --conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener" --conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener" --conf spark.sql.warehouse.dir="/user/hive/warehouse"</spark-opts>
<spark-opts>
--executor-memory ${sparkExecutorMemory}
--executor-cores ${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener"
--conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener"
--conf spark.sql.warehouse.dir="/user/hive/warehouse"
</spark-opts>
<arg>-mt</arg> <arg>yarn-cluster</arg>
<arg>-s</arg><arg>${migrationPathStep2}/all_entities</arg>
<arg>-g</arg><arg>${migrationPathStep3}</arg>
<arg>-s</arg><arg>${workingPath}/all_entities</arg>
<arg>-g</arg><arg>${graphBasePath}/graph_raw</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>