Updated on 2022-12-14 GMT+08:00

Migrating Data Between Kafka Nodes

Scenario

This section describes how to use Kafka client commands to migrate partition data between disks on a node without stopping the Kafka service.

Prerequisites

  • You have understood service requirements and prepared a Kafka user (belonging to the kafkaadmin group and not required for the normal mode).
  • The Kafka client has been installed.
  • The Kafka instance status and disk status are normal.
  • Based on the current disk space usage of the partition to be migrated, ensure that the disk space will be sufficient after the migration.

Procedure

  1. Log in as a client installation user to the node on which the Kafka client is installed.
  2. Run the following command to switch to the Kafka client installation directory, for example, /opt/kafkaclient:

    cd /opt/kafkaclient

  3. Run the following command to set environment variables:

    source bigdata_env

  4. Run the following command to authenticate the user (skip this step in normal mode):

    kinit Component service user

  5. Run the following command to switch to the Kafka client directory:

    cd Kafka/kafka/bin

  6. Run the following command to view the topic details of the partition to be migrated:

    Security mode:

    ./kafka-topics.sh --describe --bootstrap-server IP address of the Kafkacluster:21007 --command-config ../config/client.properties --topic topic name

    Normal mode:

    ./kafka-topics.sh --describe --bootstrap-server IP address of the Kafka cluster:21005 --command-config ../config/client.properties --topic Topic name

  7. Run the following command to query the mapping between Broker_ID and the IP address:

    ./kafka-broker-info.sh --zookeeper IP address of the ZooKeeper quorumpeer instance:ZooKeeper port number/kafka

    Broker_ID     IP_Address
    --------------------------
    4           192.168.0.100
    5           192.168.0.101
    6           192.168.0.102
    • IP address of the ZooKeeper quorumpeer instance

      To obtain IP addresses of all ZooKeeper quorumpeer instances, log in to FusionInsight Manager and choose Cluster > Services > ZooKeeper. On the displayed page, click Instance and view the IP addresses of all the hosts where the quorumpeer instances locate.

    • Port number of the ZooKeeper client

      Log in to FusionInsight Manager and choose Cluster > Service > ZooKeeper. On the displayed page, click Configurations and check the value of clientPort. The default value is 24002.

  8. Obtain the partition distribution and node information from the command output in 6 and 7, and create the JSON file for reallocation in the current directory.

    To migrate data in the partition whose Broker_ID is 6 to the /srv/BigData/hadoop/data1/kafka-logs directory, the required JSON configuration file is as follows:
    {"partitions":[{"topic": "testws","partition": 2,"replicas": [6,5],"log_dirs": ["/srv/BigData/hadoop/data1/kafka-logs","any"]}],"version":1}
    • topic indicates the topic name, for example, testws.
    • partition indicates the topic partition.
    • The number in replicas corresponds to Broker_ID.
    • log_dirs indicates the path of the disk to be migrated. In this example, log_dirs of the node whose Broker_ID is 5 is set to any, and that of the node whose Broker_ID is 6 is set to /srv/BigData/hadoop/data1/kafka-logs. Note that the path must correspond to the node.

  9. Run the following command to perform reallocation:

    Security mode:

    ./kafka-reassign-partitions.sh --bootstrap-server Service IP address of Broker:21007 --command-config ../config/client.properties --zookeeper {zk_host}:{port}/kafka --reassignment-json-file Path of the JSON file compiled in 8 --execute

    Normal mode:

    ./kafka-reassign-partitions.sh --bootstrap-server Service IP address of Broker:21005 --command-config ../config/client.properties --zookeeper {zk_host}:{port}/kafka --reassignment-json-file Path of the JSON file compiled in 8 --execute

    If message "Successfully started reassignment of partitions" is displayed, the execution is successful.