Does TaskExecutorPartitionHandler uses data grid? - spring-batch

Question

I am using step partitioning in my batch job which is deployed in a distributed Spring XD environment. I would like to know if TaskExecutorPartitionHandler uses data transport in our case is Rabbit MQ?
<bean id="itemReader"
class="sample.ItemReader"
scope="step" >
<property name=requestList"
value="#{stepExecutionContext[test]}" />
</bean>
<bean id="taskExecutor"
class="org.springframework.scheduling.concurrent.ThreadPoolTaskExecutor">
<property name="corePoolSize" value="10" />
<property name="maxPoolSize" value="15" />
<property name="allowCoreThreadTimeOut" value="true" />
</bean>
<batch:job id='partitionJob' restartable="false"
incrementer="jobParametersIncrementerImpl" >
<batch:step id="startLoopStep">
<batch:tasklet ref="initTasklet" />
</batch:step>
<batch:step id='partitionerStep'>
<batch:partition step="slave" partitioner="rangePartitioner">
<batch:handler grid-size="${gridSize}" task-executor="taskExecutor" />
</batch:partition>
</batch:step>
</batch:job>
<batch:step id="slave">
<batch:tasklet>
<batch:chunk reader="itemReader" writer="itemWriter"
commit-interval="1" retry-limit="3" >
</batch:chunk>
</batch:tasklet>
</batch:step>

Answer 1

No. The TaskExecutorPartitionHandler uses local threads to do the partitioning. The MessageChannelPartitionHandler is what you want for distributed partitioning. Spring XD comes with a context file that you can import for adding partitioning to single step jobs easily. Most of the out of the box jobs within Spring XD utilize this functionality and can serve as a reference.