Updated on 2022-08-12 GMT+08:00

Running a SparkSubmit Job

You can submit programs developed by yourself to MRS to execute them, and obtain the results. This section describes how to submit a Spark job on the MRS console.

Prerequisites

You have uploaded the program packages and data files required for running jobs to OBS or HDFS.

Submitting a Job on the GUI

  1. Log in to the MRS console.
  2. Choose Clusters > Active Clusters, select a running cluster, and click its name to switch to the cluster details page.
  3. If Kerberos authentication is enabled for the cluster, perform the following steps. If Kerberos authentication is not enabled for the cluster, skip this step.

    In the Basic Information area on the Dashboard page, click Synchronize on the right side of IAM User Sync to synchronize IAM users. For details, see Synchronizing IAM Users to MRS.

    • When the policy of the user group to which the IAM user belongs changes from MRS ReadOnlyAccess to MRS CommonOperations, MRS FullAccess, or MRS Administrator, wait for 5 minutes until the new policy takes effect after the synchronization is complete because the SSSD (System Security Services Daemon) cache of cluster nodes needs time to be updated. Then, submit a job. Otherwise, the job may fail to be submitted.
    • When the policy of the user group to which the IAM user belongs changes from MRS CommonOperations, MRS FullAccess, or MRS Administrator to MRS ReadOnlyAccess, wait for 5 minutes until the new policy takes effect after the synchronization is complete because the SSSD cache of cluster nodes needs time to be updated.

  4. Click the Jobs tab.
  5. Click Create. The Create Job page is displayed.
  6. Configure job information.

    Table 1 Job configuration information

    Parameter

    Description

    Name

    Job name. It contains 1 to 64 characters. Only letters, digits, hyphens (-), and underscores (_) are allowed.

    NOTE:

    You are advised to set different names for different jobs.

    Program Path

    Path of the program package to be executed. The following requirements must be met:

    • Contains a maximum of 1,023 characters, excluding special characters such as ;|&><'$. The parameter value cannot be empty or full of spaces.
    • The path of the program to be executed can be stored in HDFS or OBS. The path varies depending on the file system.
      • OBS: The path must start with obs://. Example: obs://wordcount/program/xxx.jar
      • HDFS: The path must start with /user. For details about how to import data to HDFS, see Importing Data.
    • For SparkScript and HiveScript, the path must end with .sql. For MapReduce, the path must end with .jar. For Flink and SparkSubmit, the path must end with .jar or .py. The .sql, .jar, and .py are case-insensitive.

    Program Parameter

    (Optional) Used to configure optimization parameters such as threads, memory, and vCPUs for the job to optimize resource usage and improve job execution performance.

    Table 2 describes the common parameters of a running program.

    Parameters

    (Optional) Key parameter for program execution. The parameter is specified by the function of the user's program. MRS is only responsible for loading the parameter. Multiple parameters are separated by space.

    The parameter contains a maximum of 150,000 characters. It cannot contain special characters ;|&><'$, but can be left blank.

    CAUTION:

    If you enter a parameter with sensitive information (such as the login password), the parameter may be exposed in the job details display and log printing. Exercise caution when performing this operation.

    Service Parameter

    (Optional) It is used to modify service parameters for the job. The parameter modification applies only to the current job. To make the modification take effect permanently for the cluster, follow instructions in Configuring Service Parameters.

    To add multiple parameters, click on the right. To delete a parameter, click Delete on the right.

    Table 3 lists the common service configuration parameters.

    NOTE:

    If you need to run a long-term job, such as SparkStreaming, and access OBS, you need to use Service Parameter to import the AK/SK for accessing OBS.

    Command Reference

    Command submitted to the background for execution when a job is submitted.

    Table 2 Program parameters

    Parameter

    Description

    Example Value

    --conf

    Add the task configuration items.

    spark.executor.memory=2G

    --driver-memory

    Set the running memory of driver.

    2G

    --num-executors

    Set the number of executors to be started.

    5

    --executor-cores

    Set the number of executor cores.

    2

    --class

    Set the main class of a task.

    org.apache.spark.examples.SparkPi

    --files

    Upload files to a task. The files can be custom configuration files or some data files from OBS or HDFS.

    -

    --jars

    Upload additional dependency packages of a task to add the external dependency packages to the task.

    -

    --executor-memory

    Set executor memory.

    2G

    --conf spark-yarn.maxAppAttempts

    Control the number of AM retries.

    If this parameter is set to 0, retry is not allowed. If this parameter is set to 1, one retry is allowed.

    Table 3 Service Parameter parameters

    Parameter

    Description

    Example Value

    fs.obs.access.key

    Key ID for accessing OBS.

    -

    fs.obs.secret.key

    Key corresponding to the key ID for accessing OBS.

    -

  7. Confirm job configuration information and click OK.

    After the job is created, you can manage it.

Submitting a Job in the Background

In MRS 3.x and later versions, the default installation path of the client is /opt/Bigdata/client. In MRS 3.x and earlier versions, the default installation path is /opt/client. For details, see the actual situation.

  1. Create a user for submitting jobs. For details, see Creating a User.

    In this example, a machine-machine user used in the user development scenario has been created, and user groups (hadoop and supergroup), the primary group (supergroup), and role permissions (System_administrator and default) have been correctly assigned to the user.

  2. Download the authentication credential.

    • For clusters of MRS 3.x or later, log in to FusionInsight Manager and choose System > Permission > User. In the Operation column of the newly created user, choose More > Download Authentication Credential.
    • For clusters whose version is earlier than MRS 3.x, log in to MRS Manager and choose System > Manage User. In the Operation column of the newly created user, choose More > Download Authentication Credential.

  3. Upload JAR files related to the job to the cluster. In this example, the sample JAR file built in Spark is used. It is stored in $SPARK_HOME/examples/jars/.
  4. Upload the authentication credential of the user created in 2 to the /opt/ directory of the cluster and run the following command to decompress the credential:

    tar –xvf MRSTest _xxxxxx_keytab.tar

    You will obtain two files: user.keytab and krb5.conf.

  5. Before performing operations on the cluster, run the following commands:

    source /opt/Bigdata/client/bigdata_env

    cd $SPARK_HOME

  6. Run the following command to submit the Spark job:

    ./bin/spark-submit --master yarn --deploy-mode client --conf spark.yarn.principal=MRSTest --conf spark.yarn.keytab=/opt/user.keytab --class org.apache.spark.examples.SparkPi examples/jars/spark-examples_2.11-2.3.2-mrs-2.0.jar 10

    Parameter description:

    1. Computing capability of Yarn, which specifies that the job is submitted in client mode.
    2. Configuration item of the Spark job. The authentication file and username are transferred here.
    3. spark.yarn.principal: user created in step 1
    4. spark.yarn.keytab: keytab file used for authentication
    5. xx.jar: JAR file used by the job