spark+kafka+redis统计网站访问者IP

*目的是为了防采集。需要对网站的日志信息，进行一个实时的IP访问监控。

1、kafka版本是最新的0.10.0.0

2、spark版本是1.61

3、下载对应的spark-streaming-kafka-assembly_2.10-1.6.1.jar放到spark目录下的lib目录下

下载地址 https://repo1.maven.org/maven2/org/apache/spark/

4、利用flume将nginx日志写入到kafka（后续补充）

5、编写python脚本，命名为test_spark_collect_ip.py

# coding:utf-8__author__ = 'chenhuachao''''利用pyspark连接kafka，统计访问者的IP信息，做出的一个实时的防采集'''import sysreload(sys)sys.setdefaultencoding('utf-8')import redisimport datetimefrom pyspark.streaming.kafka import KafkaUtilsfrom pyspark.streaming import StreamingContextfrom pyspark import SparkConf, SparkContextdef parse(logstring):    try:        infodict =  eval(logstring.encode('utf-8'))        ip =infodict.get('ip')        assert infodict['tj-event'] == 'onload'        assert ip        return (ip)    except:        return ()def insert_redis(rdd):    '''将符合条件的结果写入到redis'''    conn = redis.Redis(host='redis的IP',port=6380)    for i,j in rdd.collect():        print i,j        if j >=3 and j != "":            conn.sadd('cheating_ip_set_{0}'.format(datetime.datetime.now().strftime("%Y%m%d")),i)            conn.expire('cheating_ip_set',86400)if __name__ == "__main__":    topic = 'statis-detailinfo-pageevent'    sc = SparkContext(appName="pyspark_kafka_streaming_chc")    ssc = StreamingContext(sc,10)    checkpointDirectory = '/tmp/checkpoint/cp3'    ssc.checkpoint(checkpointDirectory)    kvs = KafkaUtils.createDirectStream(ssc,['statis-detailinfo-pageevent'],kafkaParams={"auto.offset.reset": "largest","metadata.broker.list":"kafka-IP:9092,kafka-IP:9092"})    #kvs.map(lambda line:line[1]).map(lambda x:parse(x)).pprint()    #这里用到了一个滑动窗口的概念，需要深入了解的可以参考http://www.kancloud.cn/kancloud/spark-programming-guide/51567    #ipcount = kvs.map(lambda line: line[1]).map(lambda x:parse(x)).map(lambda ip:(ip,1)).reduceByKey(lambda ips,num:ips+num)    ipcount = kvs.map(lambda line: line[1]).map(lambda x:parse(x)).map(lambda ip:(ip,1)).reduceByKeyAndWindow(lambda ips,num:ips+num,30,10)    # 预处理，如果需要多次计算则使用缓存    # 传入rdd进行循坏，即用于foreachRdd(insertRedis)    ipcount.foreachRDD(insert_redis)    # 各节点的rdd的循坏    # wordCounts.foreachRDD(lambda rdd: rdd.foreach(sendRecord))    ssc.start()

6、执行命令

bin/spark-submit --jars lib/spark-streaming-kafka-assembly_2.10-1.6.1.jar test_spark_collect_ip.py

7、输出界面

8、更多信息，请参考spark的官网http://spark.apache.org/docs/latest/api/python/pyspark.streaming.html#module-pyspark.streaming.kafka