Hadoop/Cascading中从FTP服务器读取数据

Question

我想从FTP服务器读取数据。我以ftp://Username:Password@host/path的格式提供驻留在FTP服务器上的文件的路径。当我使用map reduce程序从文件中读取数据时，它工作得很好。我想通过级联框架从同一文件读取数据。我正在使用层叠框架的Hfs tap来读取数据。它抛出以下异常

java.io.IOException: Stream closed
    at org.apache.hadoop.fs.ftp.FTPInputStream.close(FTPInputStream.java:98)
    at java.io.FilterInputStream.close(Unknown Source)
    at org.apache.hadoop.util.LineReader.close(LineReader.java:83)
    at org.apache.hadoop.mapred.LineRecordReader.close(LineRecordReader.java:168)
    at org.apache.hadoop.mapred.MapTask$TrackedRecordReader.close(MapTask.java:254)
    at org.apache.hadoop.mapred.MapTask.runOldMapper(MapTask.java:440)
    at org.apache.hadoop.mapred.MapTask.run(MapTask.java:372)
    at org.apache.hadoop.mapred.LocalJobRunner$Job.run(LocalJobRunner.java:212)

下面是层叠框架的代码，我从那里读取文件:

public class FTPWithHadoopDemo {
    public static void main(String args[]) {
        Tap source = new Hfs(new TextLine(new Fields("line")), "ftp://user:[e-mail]//input1");
        Tap sink = new Hfs(new TextLine(new Fields("line1")), "OP\op", SinkMode.REPLACE);
        Pipe pipe = new Pipe("First");
        pipe = new Each(pipe, new RegexSplitGenerator("\s+"));
        pipe = new GroupBy(pipe);
        Pipe tailpipe = new Every(pipe, new Count());
        FlowDef flowDef = FlowDef.flowDef().addSource(pipe, source).addTailSink(tailpipe, sink);
        new HadoopFlowConnector().connect(flowDef).complete();
    }
}

我试图在Hadoop源代码中寻找相同的异常。我发现在MapTask类中有一个处理流的方法runOldMapper。在相同的方法中，最后有一个块，其中流被关闭(in.close())。当我从finally块中删除这一行时，它工作得很好。下面是代码:

private <INKEY, INVALUE, OUTKEY, OUTVALUE> void runOldMapper(final JobConf job, final TaskSplitIndex splitIndex,
            final TaskUmbilicalProtocol umbilical, TaskReporter reporter)
                    throws IOException, InterruptedException, ClassNotFoundException {
        InputSplit inputSplit = getSplitDetails(new Path(splitIndex.getSplitLocation()), splitIndex.getStartOffset());
        updateJobWithSplit(job, inputSplit);
        reporter.setInputSplit(inputSplit);
        RecordReader<INKEY, INVALUE> in = isSkipping()
                ? new SkippingRecordReader<INKEY, INVALUE>(inputSplit, umbilical, reporter)
                : new TrackedRecordReader<INKEY, INVALUE>(inputSplit, job, reporter);
        job.setBoolean("mapred.skip.on", isSkipping());
        int numReduceTasks = conf.getNumReduceTasks();
        LOG.info("numReduceTasks: " + numReduceTasks);
        MapOutputCollector collector = null;
        if (numReduceTasks > 0) {
            collector = new MapOutputBuffer(umbilical, job, reporter);
        } else {
            collector = new DirectMapOutputCollector(umbilical, job, reporter);
        }
        MapRunnable<INKEY, INVALUE, OUTKEY, OUTVALUE> runner = ReflectionUtils.newInstance(job.getMapRunnerClass(),
                job);
        try {
            runner.run(in, new OldOutputCollector(collector, conf), reporter);
            collector.flush();
        } finally {
            // close
            in.close(); // close input
            collector.close();
        }
    }

请帮助我解决这个问题。

谢谢,Arshadali

Answer 1

经过一番努力，我发现hadoop使用org.apache.hadoop.fs.ftp.FTPFileSystem类FTP
这个类不支持查找，即从文件开始查找给定的偏移量。在一个块中读取数据，然后文件系统寻找下一个要读取的块。FTPFileSystem的默认块大小为4KB。由于不支持seek，它只能读取小于等于4KB的数据。