org.springframework.data.hadoop.mapreduce
Class JobExecutor

java.lang.Object
  org.springframework.data.hadoop.mapreduce.JobExecutor

All Implemented Interfaces:

org.springframework.beans.factory.Aware, org.springframework.beans.factory.BeanFactoryAware, org.springframework.beans.factory.DisposableBean, org.springframework.beans.factory.InitializingBean

Direct Known Subclasses:

JobRunner, JobTasklet

public abstract class JobExecutor
extends java.lang.Object
implements org.springframework.beans.factory.InitializingBean, org.springframework.beans.factory.DisposableBean, org.springframework.beans.factory.BeanFactoryAware

Common class shared for executing Hadoop Jobs.

Author:

Costin Leau, Thomas Risberg

Nested Class Summary

protected static interface

JobExecutor.JobListener

Field Summary

protected org.apache.commons.logging.Log

log

Constructor Summary

JobExecutor()

Method Summary

void	afterPropertiesSet()
void	destroy()
protected java.util.Collection<org.apache.hadoop.mapreduce.Job>	findJobs()
boolean	isKillJobsAtShutdown() Indicates whether the configured jobs should be 'killed' when the application shuts down or not.
boolean	isVerbose() Indicates whether the job execution is verbose (the default) or not.
boolean	isWaitForCompletion() Indicates whether the 'runner' should wait for the job to complete (default).
void	setBeanFactory(org.springframework.beans.factory.BeanFactory beanFactory)
void	setExecutor(java.util.concurrent.Executor executor) Sets the TaskExecutor used for executing the Hadoop job.
void	setJob(org.apache.hadoop.mapreduce.Job job) Sets the job to execute.
void	setJobNames(java.lang.String... jobName) Sets the jobs to execute by (bean) name.
void	setJobs(java.util.Collection<org.apache.hadoop.mapreduce.Job> jobs) Sets the jobs to execute.
void	setKillJobAtShutdown(boolean killJobsAtShutdown) Indicates whether the configured jobs should be 'killed' when the application shuts down (default) or not.
void	setVerbose(boolean verbose) Indicates whether the job execution is verbose (the default) or not.
void	setWaitForCompletion(boolean waitForJob) Indicates whether the 'runner' should wait for the job to complete (default) after submission or not.
protected java.util.Collection<org.apache.hadoop.mapreduce.Job>	startJobs()
protected java.util.Collection<org.apache.hadoop.mapreduce.Job>	startJobs(JobExecutor.JobListener listener)
protected java.util.Collection<org.apache.hadoop.mapreduce.Job>	stopJobs() Stops running job.
protected java.util.Collection<org.apache.hadoop.mapreduce.Job>	stopJobs(JobExecutor.JobListener listener) Stops running job.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

log

protected org.apache.commons.logging.Log log

Constructor Detail

JobExecutor

public JobExecutor()

Method Detail

afterPropertiesSet

public void afterPropertiesSet()
                        throws java.lang.Exception

Specified by:

afterPropertiesSet in interface org.springframework.beans.factory.InitializingBean

Throws:

java.lang.Exception

destroy

public void destroy()
             throws java.lang.Exception

Specified by:

destroy in interface org.springframework.beans.factory.DisposableBean

Throws:

java.lang.Exception

stopJobs

protected java.util.Collection<org.apache.hadoop.mapreduce.Job> stopJobs()

Stops running job.

Returns:

list of stopped jobs.

Throws:

java.lang.Exception

stopJobs

protected java.util.Collection<org.apache.hadoop.mapreduce.Job> stopJobs(JobExecutor.JobListener listener)

Stops running job.

Parameters:

listener - job listener

Returns:

list of stopped jobs.

Throws:

java.lang.Exception

startJobs

protected java.util.Collection<org.apache.hadoop.mapreduce.Job> startJobs()

startJobs

protected java.util.Collection<org.apache.hadoop.mapreduce.Job> startJobs(JobExecutor.JobListener listener)

findJobs

protected java.util.Collection<org.apache.hadoop.mapreduce.Job> findJobs()

setJob

public void setJob(org.apache.hadoop.mapreduce.Job job)

Sets the job to execute.

Parameters:

job - The job to execute.

setJobs

public void setJobs(java.util.Collection<org.apache.hadoop.mapreduce.Job> jobs)

Sets the jobs to execute.

Parameters:

jobs - The job to execute.

setJobNames

public void setJobNames(java.lang.String... jobName)

Sets the jobs to execute by (bean) name. This is the default method used by the hdp name space to allow lazy initialization and potential scoping to kick in.

Parameters:

jobName - The job to execute.

isWaitForCompletion

public boolean isWaitForCompletion()

Indicates whether the 'runner' should wait for the job to complete (default).

Returns:

whether to wait for the job to complete or not.

setWaitForCompletion

public void setWaitForCompletion(boolean waitForJob)

Indicates whether the 'runner' should wait for the job to complete (default) after submission or not.

Parameters:

waitForJob - whether to wait for the job to complete or not.

isVerbose

public boolean isVerbose()

Indicates whether the job execution is verbose (the default) or not.

Returns:

whether the job execution is verbose or not.

setVerbose

public void setVerbose(boolean verbose)

Indicates whether the job execution is verbose (the default) or not.

Parameters:

verbose - whether the job execution is verbose or not.

setBeanFactory

public void setBeanFactory(org.springframework.beans.factory.BeanFactory beanFactory)
                    throws org.springframework.beans.BeansException

Specified by:

setBeanFactory in interface org.springframework.beans.factory.BeanFactoryAware

Throws:

org.springframework.beans.BeansException

setExecutor

public void setExecutor(java.util.concurrent.Executor executor)

Sets the TaskExecutor used for executing the Hadoop job. By default, SyncTaskExecutor is used, meaning the calling thread is used. While this replicates the Hadoop behavior, it prevents running jobs from being killed if the application shuts down. For a fine-tuned control, a dedicated Executor is recommended.

Parameters:

executor - the task executor to use execute the Hadoop job.

isKillJobsAtShutdown

public boolean isKillJobsAtShutdown()

Indicates whether the configured jobs should be 'killed' when the application shuts down or not.

Returns:

whether or not to kill the configured jobs at shutdown

setKillJobAtShutdown

public void setKillJobAtShutdown(boolean killJobsAtShutdown)

Indicates whether the configured jobs should be 'killed' when the application shuts down (default) or not. For long-running or fire-and-forget jobs that live beyond the starting application, set this to false. Note that if setWaitForCompletion(boolean) is true, this flag is considered to be true as otherwise the application cannot shut down (since it has to keep waiting for the job).

Parameters:

killJobsAtShutdown - whether or not to kill configured jobs when the application shuts down