SparkStreaming演示

首先启动hadoop
1.sparkstreaming监听端口计算wordcount
  安装nc：yum install -y nc
  启动一个socket服务端口：nc -lk 8888
    
2.window窗口操作

3.mapWithState状态管理，实现全局累加器

4.用户行为分析系统
#在kafka中创建主题
bin/kafka-topics.sh --create --zookeeper 192.168.183.100:2181 --replication-factor 1 --partitions 3 --topic DeviceEvents

#查看kafka中已经创建的主题列表
bin/kafka-topics.sh --list --zookeeper 192.168.183.100:2181

#删除主题
bin/kafka-topics.sh --delete --zookeeper 192.168.183.100:2181 --topic DeviceEvents

#使用kafka自带的消费者客户端脚本
bin/kafka-console-consumer.sh --zookeeper 192.168.183.100:2181 --from-beginning --topic DeviceEvents

#创建mysql数据库streamdb
create database streamdb

#创建表device_click
CREATE TABLE `device_click` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`device_id` varchar(512) DEFAULT NULL,
`device_type` varchar(512) DEFAULT NULL,
`time` varchar(512) DEFAULT NULL,
`click_count` int(11) NOT NULL,
PRIMARY KEY  (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;

#kafka consumer参数
val kafkaParams = Map[String, Object](
      "bootstrap.servers" -> "192.168.183.102:9092,192.168.183.103:9092,192.168.183.104:9092",
      "key.deserializer" -> classOf[StringDeserializer],
      "value.deserializer" -> classOf[StringDeserializer],
      "group.id" -> "deviceClickstream",
      "auto.offset.reset" -> "earliest",
      "enable.auto.commi" ->  (true: java.lang.Boolean)
    )