org.apache.hadoop.mapred

Class JobClient

java.lang.Object

org.apache.hadoop.conf.Configured

org.apache.hadoop.mapreduce.tools.CLI

org.apache.hadoop.mapred.JobClient

All Implemented Interfaces:

org.apache.hadoop.conf.Configurable, org.apache.hadoop.util.Tool

@InterfaceAudience.Public
@InterfaceStability.Stable
public class JobClient
extends CLI

JobClient is the primary interface for the user-job to interact with the cluster. JobClient provides facilities to submit jobs, track their progress, access component-tasks' reports/logs, get the Map-Reduce cluster status information etc.

The job submission process involves:

1. Checking the input and output specifications of the job.
2. Computing the InputSplits for the job.
3. Setup the requisite accounting information for the DistributedCache of the job, if necessary.
4. Copying the job's jar and configuration to the map-reduce system directory on the distributed file-system.
5. Submitting the job to the cluster and optionally monitoring it's status.

Normally the user creates the application, describes various facets of the job via JobConf and then uses the JobClient to submit the job and monitor its progress.

Here is an example on how to use JobClient:

     // Create a new JobConf
     JobConf job = new JobConf(new Configuration(), MyJob.class);
     
     // Specify various job-specific parameters     
     job.setJobName("myjob");
     
     job.setInputPath(new Path("in"));
     job.setOutputPath(new Path("out"));
     
     job.setMapperClass(MyJob.MyMapper.class);
     job.setReducerClass(MyJob.MyReducer.class);

     // Submit the job, then poll for progress until the job is complete
     JobClient.runJob(job);
 

Job Control

At times clients would chain map-reduce jobs to accomplish complex tasks which cannot be done via a single map-reduce job. This is fairly easy since the output of the job, typically, goes to distributed file-system and that can be used as the input for the next job.

However, this also means that the onus on ensuring jobs are complete (success/failure) lies squarely on the clients. In such situations the various job-control options are:

1. runJob(JobConf) : submits the job and returns only after the job has completed.
2. submitJob(JobConf) : only submits the job, then poll the returned handle to the RunningJob to query status and make scheduling decisions.
3. JobConf.setJobEndNotificationURI(String) : setup a notification on job-completion, thus avoiding polling.

See Also:

JobConf, ClusterStatus, Tool, DistributedCache

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	JobClient.TaskStatusFilter

Field Summary

Fields inherited from class org.apache.hadoop.mapreduce.tools.CLI

cluster, dataPattern, headerPattern

Constructor Summary

Constructors

Constructor and Description
JobClient() Create a job client.
JobClient(org.apache.hadoop.conf.Configuration conf) Build a job client with the given Configuration, and connect to the default cluster
JobClient(InetSocketAddress jobTrackAddr, org.apache.hadoop.conf.Configuration conf) Build a job client, connect to the indicated job tracker.
JobClient(JobConf conf) Build a job client with the given JobConf, and connect to the default cluster

Method Summary

Methods

Modifier and Type	Method and Description
void	cancelDelegationToken(org.apache.hadoop.security.token.Token<DelegationTokenIdentifier> token) Deprecated. Use Token.cancel(org.apache.hadoop.conf.Configuration) instead
void	close() Close the JobClient.
void	displayTasks(JobID jobId, String type, String state) Display the information about a job's tasks, of a particular type and in a particular state
JobStatus[]	getAllJobs() Get the jobs that are submitted.
JobQueueInfo[]	getChildQueues(String queueName) Returns an array of queue information objects about immediate children of queue queueName.
TaskReport[]	getCleanupTaskReports(JobID jobId) Get the information of the current state of the cleanup tasks of a job.
Cluster	getClusterHandle() Get a handle to the Cluster
ClusterStatus	getClusterStatus() Get status information about the Map-Reduce cluster.
ClusterStatus	getClusterStatus(boolean detailed) Get status information about the Map-Reduce cluster.
protected long	getCounter(Counters cntrs, String counterGroupName, String counterName)
int	getDefaultMaps() Get status information about the max available Maps in the cluster.
int	getDefaultReduces() Get status information about the max available Reduces in the cluster.
org.apache.hadoop.security.token.Token<DelegationTokenIdentifier>	getDelegationToken(org.apache.hadoop.io.Text renewer) Get a delegation token for the user from the JobTracker.
org.apache.hadoop.fs.FileSystem	getFs() Get a filesystem handle.
RunningJob	getJob(JobID jobid) Get an RunningJob object to track an ongoing job.
RunningJob	getJob(String jobid) Deprecated. Applications should rather use getJob(JobID).
JobStatus[]	getJobsFromQueue(String queueName) Gets all the jobs which were added to particular Job Queue
TaskReport[]	getMapTaskReports(JobID jobId) Get the information of the current state of the map tasks of a job.
TaskReport[]	getMapTaskReports(String jobId) Deprecated. Applications should rather use getMapTaskReports(JobID)
org.apache.hadoop.mapred.QueueAclsInfo[]	getQueueAclsForCurrentUser() Gets the Queue ACLs for current user
JobQueueInfo	getQueueInfo(String queueName) Gets the queue information associated to a particular Job Queue
JobQueueInfo[]	getQueues() Return an array of queue information objects about all the Job Queues configured.
TaskReport[]	getReduceTaskReports(JobID jobId) Get the information of the current state of the reduce tasks of a job.
TaskReport[]	getReduceTaskReports(String jobId) Deprecated. Applications should rather use getReduceTaskReports(JobID)
JobQueueInfo[]	getRootQueues() Returns an array of queue information objects about root level queues configured
TaskReport[]	getSetupTaskReports(JobID jobId) Get the information of the current state of the setup tasks of a job.
org.apache.hadoop.fs.Path	getSystemDir() Grab the jobtracker system directory path where job-specific files are to be placed.
JobClient.TaskStatusFilter	getTaskOutputFilter() Deprecated.
static JobClient.TaskStatusFilter	getTaskOutputFilter(JobConf job) Get the task output filter out of the JobConf.
void	init(JobConf conf) Connect to the default cluster
JobStatus[]	jobsToComplete() Get the jobs that are not completed and not failed.
static void	main(String[] argv)
boolean	monitorAndPrintJob(JobConf conf, RunningJob job) Monitor a job and print status in real-time as progress is made and tasks fail.
long	renewDelegationToken(org.apache.hadoop.security.token.Token<DelegationTokenIdentifier> token) Deprecated. Use Token.renew(org.apache.hadoop.conf.Configuration) instead
static RunningJob	runJob(JobConf job) Utility that submits a job, then polls for progress until the job is complete.
void	setTaskOutputFilter(JobClient.TaskStatusFilter newValue) Deprecated.
static void	setTaskOutputFilter(JobConf job, JobClient.TaskStatusFilter newValue) Modify the JobConf to set the task output filter.
RunningJob	submitJob(JobConf conf) Submit a job to the MR system.
RunningJob	submitJob(String jobFile) Submit a job to the MR system.

Methods inherited from class org.apache.hadoop.mapreduce.tools.CLI

displayJobList, displayJobList, displayTasks, run

Methods inherited from class org.apache.hadoop.conf.Configured

getConf, setConf

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.hadoop.conf.Configurable

getConf, setConf

Constructor Detail

JobClient

public JobClient()

Create a job client.

JobClient

public JobClient(JobConf conf)
          throws IOException

Build a job client with the given JobConf, and connect to the default cluster

Parameters:

conf - the job configuration.

Throws:

IOException

JobClient

public JobClient(org.apache.hadoop.conf.Configuration conf)
          throws IOException

Build a job client with the given Configuration, and connect to the default cluster

Parameters:

conf - the configuration.

Throws:

IOException

JobClient

public JobClient(InetSocketAddress jobTrackAddr,
         org.apache.hadoop.conf.Configuration conf)
          throws IOException

Build a job client, connect to the indicated job tracker.

Parameters:

jobTrackAddr - the job tracker to connect to.

conf - configuration.

Throws:

IOException

Method Detail

init

public void init(JobConf conf)
          throws IOException

Connect to the default cluster

Parameters:

conf - the job configuration.

Throws:

IOException

close

public void close()
           throws IOException

Close the JobClient.

Throws:

IOException

getFs

public org.apache.hadoop.fs.FileSystem getFs()
                                      throws IOException

Get a filesystem handle. We need this to prepare jobs for submission to the MapReduce system.

Returns:

the filesystem handle.

Throws:

IOException

getClusterHandle

public Cluster getClusterHandle()

Get a handle to the Cluster

submitJob

public RunningJob submitJob(String jobFile)
                     throws FileNotFoundException,
                            InvalidJobConfException,
                            IOException

Submit a job to the MR system. This returns a handle to the RunningJob which can be used to track the running-job.

Parameters:

jobFile - the job configuration.

Returns:

a handle to the RunningJob which can be used to track the running-job.

Throws:

FileNotFoundException

InvalidJobConfException

IOException

submitJob

public RunningJob submitJob(JobConf conf)
                     throws FileNotFoundException,
                            IOException

Submit a job to the MR system. This returns a handle to the RunningJob which can be used to track the running-job.

Parameters:

conf - the job configuration.

Returns:

a handle to the RunningJob which can be used to track the running-job.

Throws:

FileNotFoundException

IOException

getJob

public RunningJob getJob(JobID jobid)
                  throws IOException

Get an RunningJob object to track an ongoing job. Returns null if the id does not correspond to any known job.

Parameters:

jobid - the jobid of the job.

Returns:

the RunningJob handle to track the job, null if the jobid doesn't correspond to any known job.

Throws:

IOException

getJob

@Deprecated
public RunningJob getJob(String jobid)
                  throws IOException

Deprecated. Applications should rather use getJob(JobID).

Throws:

IOException

getMapTaskReports

public TaskReport[] getMapTaskReports(JobID jobId)
                               throws IOException

Get the information of the current state of the map tasks of a job.

Parameters:

jobId - the job to query.

Returns:

the list of all of the map tips.

Throws:

IOException

getMapTaskReports

@Deprecated
public TaskReport[] getMapTaskReports(String jobId)
                               throws IOException

Deprecated. Applications should rather use getMapTaskReports(JobID)

Throws:

IOException

getReduceTaskReports

public TaskReport[] getReduceTaskReports(JobID jobId)
                                  throws IOException

Get the information of the current state of the reduce tasks of a job.

Parameters:

jobId - the job to query.

Returns:

the list of all of the reduce tips.

Throws:

IOException

getCleanupTaskReports

public TaskReport[] getCleanupTaskReports(JobID jobId)
                                   throws IOException

Get the information of the current state of the cleanup tasks of a job.

Parameters:

jobId - the job to query.

Returns:

the list of all of the cleanup tips.

Throws:

IOException

getSetupTaskReports

public TaskReport[] getSetupTaskReports(JobID jobId)
                                 throws IOException

Get the information of the current state of the setup tasks of a job.

Parameters:

jobId - the job to query.

Returns:

the list of all of the setup tips.

Throws:

IOException

getReduceTaskReports

@Deprecated
public TaskReport[] getReduceTaskReports(String jobId)
                                  throws IOException

Deprecated. Applications should rather use getReduceTaskReports(JobID)

Throws:

IOException

displayTasks

public void displayTasks(JobID jobId,
                String type,
                String state)
                  throws IOException

Display the information about a job's tasks, of a particular type and in a particular state

Parameters:

jobId - the ID of the job

type - the type of the task (map/reduce/setup/cleanup)

state - the state of the task (pending/running/completed/failed/killed)

Throws:

IOException

getClusterStatus

public ClusterStatus getClusterStatus()
                               throws IOException

Get status information about the Map-Reduce cluster.

Returns:

the status information about the Map-Reduce cluster as an object of ClusterStatus.

Throws:

IOException

getClusterStatus

public ClusterStatus getClusterStatus(boolean detailed)
                               throws IOException

Get status information about the Map-Reduce cluster.

Parameters:

detailed - if true then get a detailed status including the tracker names

Returns:

the status information about the Map-Reduce cluster as an object of ClusterStatus.

Throws:

IOException

jobsToComplete

public JobStatus[] jobsToComplete()
                           throws IOException

Get the jobs that are not completed and not failed.

Returns:

array of JobStatus for the running/to-be-run jobs.

Throws:

IOException

getAllJobs

public JobStatus[] getAllJobs()
                       throws IOException

Get the jobs that are submitted.

Returns:

array of JobStatus for the submitted jobs.

Throws:

IOException

runJob

public static RunningJob runJob(JobConf job)
                         throws IOException

Utility that submits a job, then polls for progress until the job is complete.

Parameters:

job - the job configuration.

Throws:

IOException - if the job fails

monitorAndPrintJob

public boolean monitorAndPrintJob(JobConf conf,
                         RunningJob job)
                           throws IOException,
                                  InterruptedException

Monitor a job and print status in real-time as progress is made and tasks fail.

Parameters:

conf - the job's configuration

job - the job to track

Returns:

true if the job succeeded

Throws:

IOException - if communication to the JobTracker fails

InterruptedException

setTaskOutputFilter

@Deprecated
public void setTaskOutputFilter(JobClient.TaskStatusFilter newValue)

Deprecated.

Sets the output filter for tasks. only those tasks are printed whose output matches the filter.

Parameters:

newValue - task filter.

getTaskOutputFilter

public static JobClient.TaskStatusFilter getTaskOutputFilter(JobConf job)

Get the task output filter out of the JobConf.

Parameters:

job - the JobConf to examine.

Returns:

the filter level.

setTaskOutputFilter

public static void setTaskOutputFilter(JobConf job,
                       JobClient.TaskStatusFilter newValue)

Modify the JobConf to set the task output filter.

Parameters:

job - the JobConf to modify.

newValue - the value to set.

getTaskOutputFilter

@Deprecated
public JobClient.TaskStatusFilter getTaskOutputFilter()

Deprecated.

Returns task output filter.

Returns:

task filter.

getCounter

protected long getCounter(Counters cntrs,
              String counterGroupName,
              String counterName)
                   throws IOException

Overrides:

getCounter in class CLI

Throws:

IOException

getDefaultMaps

public int getDefaultMaps()
                   throws IOException

Get status information about the max available Maps in the cluster.

Returns:

the max available Maps in the cluster

Throws:

IOException

getDefaultReduces

public int getDefaultReduces()
                      throws IOException

Get status information about the max available Reduces in the cluster.

Returns:

the max available Reduces in the cluster

Throws:

IOException

getSystemDir

public org.apache.hadoop.fs.Path getSystemDir()

Grab the jobtracker system directory path where job-specific files are to be placed.

Returns:

the system directory where job-specific files are to be placed.

getRootQueues

public JobQueueInfo[] getRootQueues()
                             throws IOException

Returns an array of queue information objects about root level queues configured

Returns:

the array of root level JobQueueInfo objects

Throws:

IOException

getChildQueues

public JobQueueInfo[] getChildQueues(String queueName)
                              throws IOException

Returns an array of queue information objects about immediate children of queue queueName.

Parameters:

queueName -

Returns:

the array of immediate children JobQueueInfo objects

Throws:

IOException

getQueues

public JobQueueInfo[] getQueues()
                         throws IOException

Return an array of queue information objects about all the Job Queues configured.

Returns:

Array of JobQueueInfo objects

Throws:

IOException

getJobsFromQueue

public JobStatus[] getJobsFromQueue(String queueName)
                             throws IOException

Gets all the jobs which were added to particular Job Queue

Parameters:

queueName - name of the Job Queue

Returns:

Array of jobs present in the job queue

Throws:

IOException

getQueueInfo

public JobQueueInfo getQueueInfo(String queueName)
                          throws IOException

Gets the queue information associated to a particular Job Queue

Parameters:

queueName - name of the job queue.

Returns:

Queue information associated to particular queue.

Throws:

IOException

getQueueAclsForCurrentUser

public org.apache.hadoop.mapred.QueueAclsInfo[] getQueueAclsForCurrentUser()
                                                                    throws IOException

Gets the Queue ACLs for current user

Returns:

array of QueueAclsInfo object for current user.

Throws:

IOException

getDelegationToken

public org.apache.hadoop.security.token.Token<DelegationTokenIdentifier> getDelegationToken(org.apache.hadoop.io.Text renewer)
                                                                                     throws IOException,
                                                                                            InterruptedException

Get a delegation token for the user from the JobTracker.

Parameters:

renewer - the user who can renew the token

Returns:

the new token

Throws:

IOException

InterruptedException

renewDelegationToken

public long renewDelegationToken(org.apache.hadoop.security.token.Token<DelegationTokenIdentifier> token)
                          throws org.apache.hadoop.security.token.SecretManager.InvalidToken,
                                 IOException,
                                 InterruptedException

Deprecated. Use Token.renew(org.apache.hadoop.conf.Configuration) instead

Renew a delegation token

Parameters:

token - the token to renew

Returns:

true if the renewal went well

Throws:

org.apache.hadoop.security.token.SecretManager.InvalidToken

IOException

InterruptedException

cancelDelegationToken

public void cancelDelegationToken(org.apache.hadoop.security.token.Token<DelegationTokenIdentifier> token)
                           throws org.apache.hadoop.security.token.SecretManager.InvalidToken,
                                  IOException,
                                  InterruptedException

Deprecated. Use Token.cancel(org.apache.hadoop.conf.Configuration) instead

Cancel a delegation token from the JobTracker

Parameters:

token - the token to cancel

Throws:

IOException

org.apache.hadoop.security.token.SecretManager.InvalidToken

InterruptedException

main

public static void main(String[] argv)
                 throws Exception

Throws:

Exception