Kafka_2.12-3.1.0 cluster environment construction
1. Environmental description
The current Kafka version still requires the use of zookeeper to store clusters, brokers, consumers and other related meta-information. The current version, namely version 2_12_3.1.0, has built-in zookeeper. : If you choose to use the
故针对zookeeper环境有两种
built
1. 使用内置的zookeeper
-in zookeeper , you need to modify the config in the Kafka installation directory. The zookeeper.properties file in the directory
2. 使用外部的zookeeper
requires the zookeeper cluster to be installed in advance.
在本案例中,我们采用了外部的zookeeper
, that is, a three-node cluster zookeeper environment was built, which was used as the metadata repository of Kafka2_12_3.1.0 version.
- The zookeeper environment configuration is as follows :
node | installation path | dataDir path |
---|---|---|
hadoop002 | /training/zookeeper-3.4.5 | /training/zookeeper-3.4.5/tmp |
hadoop003 | /training/zookeeper-3.4.5 | /training/zookeeper-3.4.5/tmp |
hadoop004 | /training/zookeeper-3.4.5 | /training/zookeeper-3.4.5/tmp |
- The kafka environment configuration is as follows :
node | installation path | Log path |
---|---|---|
hadoop002 | /training/kafka-cluster/ | /training/kafka-cluster/kafka-logs/ |
hadoop003 | /training/kafka-cluster/ | /training/kafka-cluster/kafka-logs/ |
hadoop004 | /training/kafka-cluster/ | /training/kafka-cluster/kafka-logs/ |
2. Installation steps
-
Upload the Kafka installation package, unzip and install it
tar -zvxf kafka_2.12-3.1.0.tgz -C /training
-
To modify
server.properties
the file, you only need to modify the following items:broker.id=0 log.dirs=/training/kafka-cluster/kafka-logs zookeeper.connect=hadoop002:2181,hadoop003:2181,hadoop004:2181 group.initial.rebalance.delay.ms=3
其他配置默认即可
-
Distribute the entire Kafka installation directory to other nodes
scp -r /training/kafka-cluster root@hadoop003:/training/ scp -r /training/kafka-cluster root@hadoop004:/training/
-
Go to the other two nodes and modify the server. properties to
broker.id
:1和2
,
note :broker.id 不得重复
-
Start the Kafka cluster
- Start the zookeeper cluster first (start by yourself)
- Start the Kafka cluster on hadoop002, hadoop003, and hadoop004 in sequence, and execute the following command:
kafka-server-start.sh server.properties &
注意:该命令是相对路径下执行且以后台进程方式启动Kafka
- Verification: Execute:
jps
command on three nodes and you will see that the process name is:Kafka
Kafka_2.12-3.1.0 cluster environment construction
1. Environmental description
The current Kafka version still requires the use of zookeeper to store clusters, brokers, consumers and other related meta-information. The current version, namely version 2_12_3.1.0, has built-in zookeeper. : If you choose to use the
故针对zookeeper环境有两种
built
1. 使用内置的zookeeper
-in zookeeper , you need to modify the config in the Kafka installation directory. The zookeeper.properties file in the directory
2. 使用外部的zookeeper
requires the zookeeper cluster to be installed in advance.
在本案例中,我们采用了外部的zookeeper
, that is, a three-node cluster zookeeper environment was built, which was used as the metadata repository of Kafka2_12_3.1.0 version.
- The zookeeper environment configuration is as follows :
node | installation path | dataDir path |
---|---|---|
hadoop002 | /training/zookeeper-3.4.5 | /training/zookeeper-3.4.5/tmp |
hadoop003 | /training/zookeeper-3.4.5 | /training/zookeeper-3.4.5/tmp |
hadoop004 | /training/zookeeper-3.4.5 | /training/zookeeper-3.4.5/tmp |
- The kafka environment configuration is as follows :
node | installation path | Log path |
---|---|---|
hadoop002 | /training/kafka-cluster/ | /training/kafka-cluster/kafka-logs/ |
hadoop003 | /training/kafka-cluster/ | /training/kafka-cluster/kafka-logs/ |
hadoop004 | /training/kafka-cluster/ | /training/kafka-cluster/kafka-logs/ |
2. Installation steps
-
Upload the Kafka installation package, unzip and install it
tar -zvxf kafka_2.12-3.1.0.tgz -C /training
-
To modify
server.properties
the file, you only need to modify the following items:broker.id=0 log.dirs=/training/kafka-cluster/kafka-logs zookeeper.connect=hadoop002:2181,hadoop003:2181,hadoop004:2181 group.initial.rebalance.delay.ms=3
其他配置默认即可
-
Distribute the entire Kafka installation directory to other nodes
scp -r /training/kafka-cluster root@hadoop003:/training/ scp -r /training/kafka-cluster root@hadoop004:/training/
-
Go to the other two nodes and modify the server. properties to
broker.id
:1和2
,
note :broker.id 不得重复
-
Start the Kafka cluster
- Start the zookeeper cluster first (start by yourself)
- Start the Kafka cluster on hadoop002, hadoop003, and hadoop004 in sequence, and execute the following command:
kafka-server-start.sh server.properties &
注意:该命令是相对路径下执行且以后台进程方式启动Kafka
- Verification: Execute:
jps
command on three nodes and you will see that the process name is:Kafka