ExecutorRunner¶

Creating Instance¶

ExecutorRunner takes the following to be created:

• Application ID
• Executor ID
• ApplicationDescription
• Number of CPU cores
• Amount of Memory
• RpcEndpointRef
• Worker ID
• web UI's scheme
• Host
• web UI's port
• Public Address
• Spark's Home Directory
• Executor's Directory
• Worker's URL
• SparkConf
• Local Directories of the Spark Application
• Executor State
• Map[String, ResourceInformation] (default: empty)

ExecutorRunner is created when:

• Worker is requested to handle a LaunchExecutor message

Starting ExecutorRunner¶

start(): Unit

start...FIXME

start is used when:

• Worker is requested to handle a LaunchExecutor message

fetchAndRunExecutor¶

fetchAndRunExecutor(): Unit

fetchAndRunExecutor...FIXME

killProcess¶

killProcess(
  message: Option[String]): Unit

killProcess...FIXME

killProcess is used when:

• ExecutorRunner is requested to fetchAndRunExecutor (and fails) and shut down