安装Flink1.9
1、准备:jdk 1.8
SSH免密登录
三台以上机器(我准备了五台):centos7-1、centos7-2、centos7-3、centos7-4、centos7-5
zookeeper: centos7-3、centos7-4、centos7-5
...
2、从官网下载flink:https://flink.apache.org/downloads.html
3、把文件tar开
$ tar -zxvf flink-1.9.0-bin-scala_2.11.tgz
4、 修改配置
$ vim /conf/slaves
#添加一下内容
centos7-2
centos7-3
centos7-4
centos7-5
$ vim /conf/flink-conf.yaml
jobmanager.rpc.address: 192.168.2.141
5、启动Flink集群
$ bin/start-cluster.sh
6、验证
# centos7-1机器
$ jps
18376 Jps
14922 StandaloneSessionClusterEntrypoint
# 其他机器
$ jps
9899 Jps
9821 TaskManagerRunner
7、Flink的web ui
http://localhost:8081 (localhost替换成实际的ip)
JobManager高可用配置(HA)
1、编辑masters文件
$ vim /conf/masters
centos7-1:8081
centos7-2:8081
centos7-3:8081
2、编辑flink-conf.yaml
$ vim /conf/flink-conf.yaml
high-availability: zookeeper
high-availability.storageDir: hdfs://centos7-1:8020/flink/ha/
high-availability.zookeeper.quorum: centos7-3:2181,centos7-4:2181,centos7-5:2181
high-availability.zookeeper.path.root: /flink
3、编辑zoo.cfg
$ vim /conf/zoo.cfg
server.1=centos7-3:2888:3888
server.2=centos7-4:2888:3888
server.3=centos7-5:2888:3888
4、分发到其他机器
$ scp -r flink-1.9.0 hadoop@centos7-2:/home/hadoop/
$ scp -r flink-1.9.0 hadoop@centos7-3:/home/hadoop/
$ scp -r flink-1.9.0 hadoop@centos7-4:/home/hadoop/
$ scp -r flink-1.9.0 hadoop@centos7-5:/home/hadoop/
5、启动集群
$ ./bin/start-cluster.sh
Starting HA cluster with 3 masters.
Starting standalonesession daemon on host centos7-1.
Starting standalonesession daemon on host centos7-2.
Starting standalonesession daemon on host centos7-3.
Starting taskexecutor daemon on host centos7-2.
Starting taskexecutor daemon on host centos7-3.
Starting taskexecutor daemon on host centos7-4.
Starting taskexecutor daemon on host centos7-5.
6、查看进程
#masters文件配置的机器:StandaloneSessionClusterEntrypoint
#slaves文件配置的机器:TaskManagerRunner
$ jps
15908 StandaloneSessionClusterEntrypoint
16151 TaskManagerRunner
7、Flink的web ui
http://localhost:8081 (localhost替换成实际的ip)
简单使用