Configuring Flink - Amazon EMR
Parallelism optionsConfigurable filesConfiguring Flink on an EMR Cluster with multiple master nodesConfiguring memory process sizeConfiguring log output file size
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China.

Configuring Flink

You may want to configure Flink using a configuration file. For example, the main configuration file for Flink is called flink-conf.yaml. This is configurable using the Amazon EMR configuration API.

To configure the number of task slots used for Flink using the Amazon CLI

  1. Create a file, configurations.json, with the following content:

    [
    {
      "Classification": "flink-conf",
      "Properties": {
        "taskmanager.numberOfTaskSlots":"2"
      }
    }
]

  2. Next, create a cluster with the following configuration:

    aws emr create-cluster --release-label emr-5.36.0 \
--applications Name=Flink \
--configurations file://./configurations.json \
--region us-east-1 \
--log-uri s3://myLogUri \
--instance-type m5.xlarge \
--instance-count 2 \
--service-role EMR_DefaultRole \ 
--ec2-attributes KeyName=YourKeyName,InstanceProfile=EMR_EC2_DefaultRole
Note

It is also possible to change some configurations using the Flink API. For more information, see Concepts in the Flink documentation.

With Amazon EMR version 5.21.0 and later, you can override cluster configurations and specify additional configuration classifications for each instance group in a running cluster. You do this by using the Amazon EMR console, the Amazon Command Line Interface (Amazon CLI), or the Amazon SDK. For more information, see Supplying a Configuration for an Instance Group in a Running Cluster.

As the owner of your application, you know best what resources should be assigned to tasks within Flink. For the purposes of the examples in this documentation, use the same number of tasks as the slave instances that you use for the application. We generally recommend this for the initial level of parallelism but you can also increase the granularity of parallelism using task slots, which should generally not exceed the number of virtual cores per instance. For more information about Flink's architecture, see Concepts in the Flink documentation.

Currently, the files that are configurable within the Amazon EMR configuration API are:

  • flink-conf.yaml

  • log4j.properties

  • flink-log4j-session

  • log4j-cli.properties

The JobManager of Flink remains available during the master node failover process in an EMR cluster with multiple master nodes. Beginning with Amazon EMR version 5.28.0, JobManager high availability is also enabled automatically. No manual configuration is needed.

With Amazon EMR versions 5.27.0 or earlier, the JobManager is a single point of failure. When the JobManager fails, it loses all job states and will not resume the running jobs. You can enable JobManager high availability by configuring application attempt count, checkpointing, and enabling ZooKeeper as state storage for Flink, as the following example demonstrates:

[
  {
    "Classification": "yarn-site",
    "Properties": {
      "yarn.resourcemanager.am.max-attempts": "10"
    }
  },
  {
    "Classification": "flink-conf",
    "Properties": {
        "yarn.application-attempts": "10",
        "high-availability": "zookeeper",
        "high-availability.zookeeper.quorum": "%{hiera('hadoop::zk')}",
        "high-availability.storageDir": "hdfs:///user/flink/recovery",
        "high-availability.zookeeper.path.root": "/flink"
    }
  }
]

You must configure both maximum application master attempts for YARN and application attempts for Flink. For more information, see Configuration of YARN cluster high availability. You may also want to configure Flink checkpointing to make restarted JobManager recover running jobs from previously completed checkpoints. For more information, see Flink checkpointing.

For Amazon EMR versions that use Flink 1.11.x, you must configure the total memory process size for both JobManager (jobmanager.memory.process.size) and TaskManager (taskmanager.memory.process.size) in flink-conf.yaml. You can set these values by either configuring the cluster with the configuration API or manually uncommenting these fields via SSH. Flink provides the following default values.

  • jobmanager.memory.process.size: 1600m

  • taskmanager.memory.process.size: 1728m

To exclude JVM metaspace and overhead, use the total Flink memory size (taskmanager.memory.flink.size) instead of taskmanager.memory.process.size. The default value for taskmanager.memory.process.size is 1280m. It's not recommended to set both taskmanager.memory.process.size and taskmanager.memory.process.size.

All Amazon EMR versions using Flink 1.12.0 and later have the default values listed in Flink's open-source set as the default values on Amazon EMR, so you don't need to configure them yourself.

Flink application containers create and write to three types of log files: .out files, .log files, and .err files. Only .err files are compressed and removed from the file system, while .log and .out log files remain in the file system. To ensure these output files remain manageable and the cluster remains stable, you can configure log rotation in log4j.properties to set a maximum number of files and limit their sizes.

Amazon EMR versions 5.30.0 and later

Starting with Amazon EMR 5.30.0, Flink uses the log4j2 logging framework with the configuration classification name flink-log4j. The following example configuration demonstrates the log4j2 format.

[
  {
    "Classification": "flink-log4j",
    "Properties": {
      "appender.rolling.name": "RollingFileAppender",
      "appender.rolling.type":"RollingFile",
      "appender.rolling.append" : "false",
      "appender.rolling.fileName" : "${sys:log.file}",
      "appender.rolling.filePattern" : "${sys:log.file}.%i",
      "appender.rolling.layout.type" : "PatternLayout",
      "appender.rolling.layout.pattern" : "%d{yyyy-MM-dd HH:mm:ss,SSS} %-5p %-60c %x - %m%n",
      "appender.rolling.policies.type" : "Policies",
      "appender.rolling.policies.size.type" : "SizeBasedTriggeringPolicy",
      "appender.rolling.policies.size.size" : "100MB",
      "appender.rolling.strategy.type" : "DefaultRolloverStrategy",
      "appender.rolling.strategy.max" : "10"
    },
  }
]

Amazon EMR versions 5.29.0 and earlier

With Amazon EMR versions 5.29.0 and earlier, Flink uses the log4j logging framework. The following example configuration demonstrates the log4j format.

[
  {
    "Classification": "flink-log4j",
    "Properties": {
      "log4j.appender.file": "org.apache.log4j.RollingFileAppender",
      "log4j.appender.file.append":"true",
      # keep up to 4 files and each file size is limited to 100MB
      "log4j.appender.file.MaxFileSize":"100MB",
      "log4j.appender.file.MaxBackupIndex":4,
      "log4j.appender.file.layout":"org.apache.log4j.PatternLayout",
      "log4j.appender.file.layout.ConversionPattern":"%d{yyyy-MM-dd HH:mm:ss,SSS} %-5p %-60c %x - %m%n"
    },
  }
]