ไธญ

Spark Node

Overview

Spark task type used to execute Spark program. For Spark nodes, the worker submits the task by using the spark command spark submit. See spark-submit for more details.

Create Task

  • Click Project -> Management-Project -> Name-Workflow Definition, and click the Create Workflow button to enter the DAG editing page.
  • Drag from the toolbar to the canvas.

Task Parameter

  • Node name: The node name in a workflow definition is unique.
  • Run flag: Identifies whether this node schedules normally, if it does not need to execute, select the prohibition execution.
  • Descriptive information: Describe the function of the node.
  • Task priority: When the number of worker threads is insufficient, execute in the order of priority from high to low, and tasks with the same priority will execute in a first-in first-out order.
  • Worker grouping: Assign tasks to the machines of the worker group to execute. If Default is selected, randomly select a worker machine for execution.
  • Environment Name: Configure the environment name in which run the script.
  • Times of failed retry attempts: The number of times the task failed to resubmit.
  • Failed retry interval: The time interval (unit minute) for resubmitting the task after a failed task.
  • Delayed execution time: The time (unit minute) that a task delays in execution.
  • Timeout alarm: Check the timeout alarm and timeout failure. When the task runs exceed the "timeout", an alarm email will send and the task execution will fail.
  • Program type: Supports Java, Scala and Python.
  • Spark version: Support Spark1 and Spark2.
  • The class of main function: The full path of Main Class, the entry point of the Spark program.
  • Main jar package: The Spark jar package (upload by Resource Center).
  • Deployment mode: Support 3 deployment modes: yarn-cluster, yarn-client and local.
  • Task name (optional): Spark task name.
  • Driver core number: Set the number of Driver core, which can be set according to the actual production environment.
  • Driver memory size: Set the size of Driver memories, which can be set according to the actual production environment.
  • Number of Executor: Set the number of Executor, which can be set according to the actual production environment.
  • Executor memory size: Set the size of Executor memories, which can be set according to the actual production environment.
  • Main program parameters: Set the input parameters of the Spark program and support the substitution of custom parameter variables.
  • Optional parameters: support --jars, --files,--archives, --conf format.
  • Resource: Appoint resource files in the Resource if parameters refer to them.
  • Custom parameter: It is a local user-defined parameter for Spark, and will replace the content with ${variable} in the script.
  • Predecessor task: Selecting a predecessor task for the current task, will set the selected predecessor task as upstream of the current task.

Task Example

Execute the WordCount Program

This is a common introductory case in the big data ecosystem, which often apply to computational frameworks such as MapReduce, Flink and Spark. The main purpose is to count the number of identical words in the input text. (Flink's releases attach this example job)

Configure the Spark Environment in DolphinScheduler

If you are using the Spark task type in a production environment, it is necessary to configure the required environment first. The following is the configuration file: worker/bin/dolphinscheduler_env.sh.

spark_configure

Upload the Main Package

When using the Spark task node, you need to upload the jar package to the Resource Centre for the execution, refer to the resource center.

After finish the Resource Centre configuration, upload the required target files directly by dragging and dropping.

resource_upload

Configure Spark Nodes

Configure the required content according to the parameter descriptions above.

demo-spark-simple

Notice

JAVA and Scala only used for identification, there is no difference. If you use Python to develop Spark application, there is no class of the main function and the rest is the same.