flume kafak to hdfs

#example.conf: A single-node Flume configuration

#Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1

#Describe/configure the source
a1.sources.r1.type = org.apache.flume.source.kafka.KafkaSource
a1.sources.r1.kafka.bootstrap.servers = hdp4:6667,hdp5:6667,hdp6:6667
a1.sources.r1.kafka.consumer.group.id = flume_test
# 定义kafka所在zk的地址
a1.sources.r1.zookeeperConnect = hdp1.hdp:2181,hdp2.hdp:2181,hdp3.hdp:2181
a1.sources.r1.kafka.topics = IOT_DS_DATA_BACK2
# 消费超时时间
a1.sources.r1.kafka.consumer.timeout.ms = 10000
#向channel写入消息的最多条数
a1.sources.r1.batchSize = 50000
#向channel书写的最大时间 (毫秒)
a1.sources.r1.batchDurationMillis = 2000
#9.7mb
a1.sources.r1.kafka.consumer.max.partition.fetch.bytes = 10240000
# Describe the sink
a1.sinks.k1.type = hdfs
# hdfs://hdp1.hdp:8020/IOT/*/%y-%m-%d/%H-%M
a1.sinks.k1.hdfs.path = hdfs://hdp1.hdp:8020/IOT/*/20%y/%m/%d-%H-%M
#用本地时间格式化目录
a1.sinks.k1.hdfs.useLocalTimeStamp = true
#下沉后, 生成的文件类型，默认是Sequencefile，可用DataStream，则为普通文本  a1.sinks.k1.hdfs.fileType = DataStream
a1.sinks.k1.hdfs.writeFormat = Text
a1.sinks.k1.hdfs.fileType = DataStream
#fileType: 默认值为SequenceFile，文件格式主要包括：SequenceFile, DataStream, CompressedStream；当使用DataStream时，文件不会被压缩，则不需要设置hdfs.codeC; 当使用CompressedStream时，则必须设置一个正确的hdfs.codeC值
#a1.sinks.k1.hdfs.fileType = CompressedStream
#codeC: 文件压缩格式，包括：gzip, bzip2, lzo, lzop, snappy;
#a1.sinks.k1.hdfs.codeC = lzop

#前缀
a1.sinks.k1.hdfs.filePrefix = data
#后缀
a1.sinks.k1.hdfs.fileSuffix=.json
#文件回滚之前等待的时时间单位 秒
a1.sinks.k1.hdfs.rollInterval = 1800
#文件滚动大的大小限制（bytes）8000000  /1024/1024
a1.sinks.k1.hdfs.rollSize =  90000000
#写入多少个 event 数据后滚动文件（事件个数）
a1.sinks.k1.hdfs.rollCount = 70000
#hdfs sink启动的操作HDFS的线程数。
a1.sinks.k1.hdfs.threadsPoolSize = 100
#默认值：1，hdfs sink 启动的根据时间滚动文件的线程数。
a1.sinks.k1.hdfs.rollTimerPoolSize = 10
# 事件就往里面写入
a1.sinks.k1.hdfs.batchSize = 10000
a1.sinks.k1.hdfs.txnEventMax = 10000

# 10 分钟就创建文件
a1.sinks.k1.hdfs.round = true
a1.sinks.k1.hdfs.roundValue = 10
a1.sinks.k1.hdfs.roundUnit = minute
#默认值为0, 当目前被打开的临时文件在该参数指定的时间（秒）内，没有任何数据写入，则将该临时文件关闭并重命名成目标文件
#a1.sinks.k1.hdfs.idleTimeout=300
#默认值为HDFS副本数；写入HDFS文件块的最小副本数，该参数会影响文件的滚动配置，一般将该参数配置成1,才可以按照配置正确滚动文件；
a1.sinks.k1.hdfs.minBlockReplicas=1
#默认值5000,最大允许打开的HDFS文件数，当打开的文件数达到该值，最早打开的文件将会被关闭；
a1.sinks.k1.hdfs.maxOpenFiles=5000
#默认值：10000，执行HDFS操作的超时时间（单位：毫秒）；
a1.sinks.k1.hdfs.callTimeout=100000

# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 800000000
a1.channels.c1.transactionCapacity = 500000
a1.channels.c1.keep-alive = 50

# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1
      private String filePathUrl;
                       

 public Status process() throws EventDeliveryException {
    Channel channel = getChannel();
    Transaction transaction = channel.getTransaction();
    transaction.begin();
    try {
      Set<BucketWriter> writers = new LinkedHashSet<>();
      int txnEventCount = 0;
      for (txnEventCount = 0; txnEventCount < batchSize; txnEventCount++) {
        Event event = channel.take();
        if (event == null) {
          break;
        }
        //自定义-----------------------
        byte[] body = event.getBody();
        String msg = new String(body);
        //System.out.println("event.getBody().toString():"+msg);
        Map data = parseData(msg);
        String date = data.get("t").toString();
        String orgId = data.get("orgId").toString();

        filePath="hdfs://hdp1.hdp:8020/IOT/"+newInteger(orgId)%100+"/"+orgId+"/"+filePathUrl.substring(filePathUrl.indexOf("*")+2,filePathUrl.length());

        //自定义----------------------------------
        // reconstruct the path name by substituting place holders
        String realPath = BucketPath.escapeString(filePath, event.getHeaders(),
            timeZone, needRounding, roundUnit, roundValue, useLocalTime);
        String realName = BucketPath.escapeString(fileName, event.getHeaders(),
            timeZone, needRounding, roundUnit, roundValue, useLocalTime);

        String lookupPath = realPath + DIRECTORY_DELIMITER + realName;
        BucketWriter bucketWriter;
        HDFSWriter hdfsWriter = null;
        // Callback to remove the reference to the bucket writer from the
        // sfWriters map so that all buffers used by the HDFS file
        // handles are garbage collected.
        WriterCallback closeCallback = new WriterCallback() {
          @Override
          public void run(String bucketPath) {