Updated on 2022-07-11 GMT+08:00

Scenarios

Scenarios

Assume that a Flink service receives one word record every 1 second.

Develop a Flink application that can generate output of prefixed message contents.

Data Planning

Sample project data of Flink is stored in Kafka. A user with Kafka permission can send data to Kafka and receive data from it.

  1. Ensure that clusters, including HDFS, YARN, Flink, and Kafka are successfully installed.
  2. Create a topic.

    The format of the command is following:

    bin/kafka-topics.sh --create --zookeeper {zkQuorum}/kafka --partitions {partitionNum} --replication-factor {replicationNum} --topic {Topic}

    Table 1

    parameter

    Description

    {zkQuorum}

    ZooKeeper cluster information. The format is IP:port.

    {PartitionNum}

    The number of partitions for the topic.

    {ReplicationNum}

    The number of copies of each partition for the topic.

    {Topic}

    The topic name.

    Assume that the IP:ports of ZooKeeper clusters are 10.96.101.32:2181, 10.96.101.251:2181, 10.96.101.177:2181, and 10.91.8.160:2181, and the topic named is topic1. The command for creating a topic is as follows:
    bin/kafka-topics.sh --create --zookeeper 10.96.101.32:2181,10.96.101.251:2181,10.96.101.177:2181,10.91.8.160:2181/kafka --partitions 5 --replication-factor 1 --topic topic1

Development Approach

  1. Start the Flink Kafka Producer to send data to Kafka.
  2. Start Flink Kafka Consumer to receive data from Kafka. Ensure that topics of Kafka Consumer are consistent with that of Kafka Producer.
  3. Add prefix to the data content and print the result.