This topic takes E-MapReduce-Flume moving audit log to HDFS as an example to describe how to use Flume.

Background

E-MapReduce supports cluster management for EMR-Flume since V3.19.0. By using cluster management, you can configure and manage Flume Agent on the web pages. In the example, Flume Agent is started on the master instance to collect audit log stored in the local disk and sends the log to core instances by using the Avro protocol. Failover Sink Processor is configured and started on the core instances to receive the data from the master instance. Then the data is moved to to HDFS by using sinks. For more scenarios of using Flume, see Configure Flume.

Preparations

Create an EMR cluster and select Flume from the optional services. For more information, see Create an E-MapReduce cluster.

Procedure

  • Configure and start Flume Agent on core instances
    For example, configure Flume Agent for the core instance group on the emr-worker-1 node as shown in the following figure.
    1. In the Service Configuration section, set the configurations as follows.
      default-agent.sinks.default-sink.type hdfs
      default-agent.channels.default-channel.type file
      default-agent.sources.default-source.type avro
      deploy_node_hostname emr-worker-1
    2. Click Custom Configuration and add the following configurations.
      default-agent.sinks.default-sink.hdfs.path For high-availability clusters, the hdfs://emr-cluster/path format is used for the address.
      default-agent.sinks.default-sink.hdfs.fileType DataStream
      default-agent.sinks.default-sink.hdfs.rollSize 0
      default-agent.sinks.default-sink.hdfs.rollCount 0
      default-agent.sinks.default-sink.hdfs.rollInterval 86400
      default-agent.sinks.default-sink.hdfs.batchSize 51200
      default-agent.sources.default-source.bind 0.0.0.0
      default-agent.sources.default-source.port Set a value as required.
      default-agent.channels.default-channel.transactionCapacity 51200
      default-agent.channels.default-channel.dataDirs The path on which channels store the event data.
      Default-agent.channels.default-channel.checkpointDir The path on which the checkpoint file is stored.
      default-agent.channels.default-channel.capacity Set a value based on the hdfs roll.
    3. Save the configuration and start Flume Agent.
    4. Click History. After Successful appears in the Status column, STARTED is displayed for Flume Agent on the emr-worker-1 node in the Status column in the Component Deployment tab page.After Flume Agent on the emr-worker-1 node is started, start Flume Agent on other worker nodes. For example, to start Flume Agent on the emr-worker-2 node, modify the following configuration items.
      deploy_node_hostname The hostname of the node.
      default-agent.sinks.default-sink.hdfs.path For high-availability clusters, the hdfs://emr-cluster/path format is used for the address.
    5. Save the configuration, start all components, and select emr-worker-2 as the target node.
  • Configure and start Flume Agent on the master instance

    For example,

    Configure Agent as follows.
    additional_sinks k1
    deploy_node_hostname emr-header-1
    default-agent.sources.default-source.type taildir
    default-agent.sinks.default-sink.type avro
    default-agent.channels.default-channel.type file
    The new configurations are as follows.
    Configuration item Value
    default-agent.sources.default-source.filegroups f1
    default-agent.sources.default-source.filegroups.f1 /mnt/disk1/log/hadoop-hdfs/hdfs-audit.log. *
    default-agent.sources.default-source.positionFile The path on which the position file is stored.
    default-agent.channels.default-channel.checkpointDir The path on which the checkpoint file is stored.
    default-agent.channels.default-channel.dataDirs The path on which channels store the event data.
    default-agent.channels.default-channel.capacity Set a value as required.
    default-agent.sources.default-source.batchSize 2000
    default-agent.channels.default-channel.transactionCapacity 2000
    default-agent.sources.default-source.ignoreRenameWhenMultiMatching true
    default-agent.sinkgroups g1
    default-agent.sinkgroups.g1.sinks default-sink k1
    default-agent.sinkgroups.g1.processor.type failover
    default-agent.sinkgroups.g1.processor.priority.default-sink 10
    default-agent.sinkgroups.g1.processor.priority.k1 5
    default-agent.sinks.default-sink.hostname The IP address of the emr-worker-1 node.
    default-agent.sinks.default-sink.port The port of Flume Agent on the emr-worker-1 node.
    default-agent.sinks.k1.hostname The IP address of the emr-worker- node.
    default-agent.sinks.k1.port The port of Flume Agent on the emr-worker-2 node.
    default-agent.sinks.default-sink.batch-size 2000
    default-agent.sinks.k1.batch-size 2000
    default-agent.sinks.k1.type avro
    default-agent.sinks.k1.channel default-channel

View the result of synchronization

By using the HDFS command, you can see that the data is written to files named FlumeData.${timestamp}. "timestamp" shows when the file was created.