Run MapReduce jobs with Apache Hadoop on HDInsight using PowerShell

บทความ
06/15/2024

This document provides an example of using Azure PowerShell to run a MapReduce job in a Hadoop on HDInsight cluster.

Prerequisites

An Apache Hadoop cluster on HDInsight. See Create Apache Hadoop clusters using the Azure portal.
The PowerShell Az Module installed.

Run a MapReduce job

Azure PowerShell provides cmdlets that allow you to remotely run MapReduce jobs on HDInsight. Internally, PowerShell makes REST calls to WebHCat (formerly called Templeton) running on the HDInsight cluster.

The following cmdlets used when running MapReduce jobs in a remote HDInsight cluster.

Cmdlet	Description
Connect-AzAccount	Authenticates Azure PowerShell to your Azure subscription.
New-AzHDInsightMapReduceJobDefinition	Creates a new job definition by using the specified MapReduce information.
Start-AzHDInsightJob	Sends the job definition to HDInsight and starts the job. A job object is returned.
Wait-AzHDInsightJob	Uses the job object to check the status of the job. It waits until the job completes or the wait time is exceeded.
Get-AzHDInsightJobOutput	Used to retrieve the output of the job.

The following steps demonstrate how to use these cmdlets to run a job in your HDInsight cluster.

Using an editor, save the following code as mapreducejob.ps1.

# Login to your Azure subscription
$context = Get-AzContext
if ($context -eq $null) 
{
    Connect-AzAccount
}
$context

# Get cluster info
$clusterName = Read-Host -Prompt "Enter the HDInsight cluster name"
$creds=Get-Credential -Message "Enter the login for the cluster"

#Get the cluster info so we can get the resource group, storage, etc.
$clusterInfo = Get-AzHDInsightCluster -ClusterName $clusterName
$resourceGroup = $clusterInfo.ResourceGroup
$storageAccountName=$clusterInfo.DefaultStorageAccount.split('.')[0]
$container=$clusterInfo.DefaultStorageContainer
#NOTE: This assumes that the storage account is in the same resource
#      group as the cluster. If it is not, change the
#      --ResourceGroupName parameter to the group that contains storage.
$storageAccountKey=(Get-AzStorageAccountKey `
    -Name $storageAccountName `
-ResourceGroupName $resourceGroup)[0].Value

#Create a storage context
$context = New-AzStorageContext `
    -StorageAccountName $storageAccountName `
    -StorageAccountKey $storageAccountKey

#Define the MapReduce job
#NOTE: If using an HDInsight 2.0 cluster, use hadoop-examples.jar instead.
# -JarFile = the JAR containing the MapReduce application
# -ClassName = the class of the application
# -Arguments = The input file, and the output directory
$wordCountJobDefinition = New-AzHDInsightMapReduceJobDefinition `
    -JarFile "/example/jars/hadoop-mapreduce-examples.jar" `
    -ClassName "wordcount" `
    -Arguments `
        "/example/data/gutenberg/davinci.txt", `
        "/example/data/WordCountOutput"

#Submit the job to the cluster
Write-Host "Start the MapReduce job..." -ForegroundColor Green
$wordCountJob = Start-AzHDInsightJob `
    -ClusterName $clusterName `
    -JobDefinition $wordCountJobDefinition `
    -HttpCredential $creds

#Wait for the job to complete
Write-Host "Wait for the job to complete..." -ForegroundColor Green
Wait-AzHDInsightJob `
    -ClusterName $clusterName `
    -JobId $wordCountJob.JobId `
    -HttpCredential $creds
# Download the output
Get-AzStorageBlobContent `
    -Blob 'example/data/WordCountOutput/part-r-00000' `
    -Container $container `
    -Destination output.txt `
    -Context $context
# Print the output of the job.
Get-AzHDInsightJobOutput `
    -Clustername $clusterName `
    -JobId $wordCountJob.JobId `
    -HttpCredential $creds

Open a new Azure PowerShell command prompt. Change directories to the location of the mapreducejob.ps1 file, then use the following command to run the script:
```
.\mapreducejob.ps1
```
When you run the script, you're prompted for the name of the HDInsight cluster and the cluster login. You may also be prompted to authenticate to your Azure subscription.
When the job completes, you receive output similar to the following text:
```
Cluster         : CLUSTERNAME
ExitCode        : 0
Name            : wordcount
PercentComplete : map 100% reduce 100%
Query           :
State           : Completed
StatusDirectory : f1ed2028-afe8-402f-a24b-13cc17858097
SubmissionTime  : 12/5/2014 8:34:09 PM
JobId           : job_1415949758166_0071
```
This output indicates that the job completed successfully.

Note

If the ExitCode is a value other than 0, see Troubleshooting.

This example also stores the downloaded files to an output.txt file in the directory that you run the script from.

View output

To see the words and counts produced by the job, open the output.txt file in a text editor.

Note

The output files of a MapReduce job are immutable. So if you rerun this sample, you need to change the name of the output file.

Troubleshooting

If no information is returned when the job completes, view errors for the job. To view error information for this job, add the following command to the end of the mapreducejob.ps1 file. Then save the file and rerun the script.

# Print the output of the WordCount job.
Write-Host "Display the standard output ..." -ForegroundColor Green
Get-AzHDInsightJobOutput `
        -Clustername $clusterName `
        -JobId $wordCountJob.JobId `
        -HttpCredential $creds `
        -DisplayOutputType StandardError

This cmdlet returns the information that was written to STDERR as the job runs.

Next steps

As you can see, Azure PowerShell provides an easy way to run MapReduce jobs on an HDInsight cluster, monitor the job status, and retrieve the output. For information about other ways you can work with Hadoop on HDInsight:

แชร์ผ่าน