Flume实战监听文件夹内文件变化
Flume官网有多种场景的source，sink，channel的配置
1、flume安装目录下新建文件夹 example
2、在example下新建文件
spooldir-logger.conf内容如下：
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = spooldir
a1.sources.r1.spoolDir = /home/hadoop/flume_test
a1.sources.r1.fileHeader = true
# Describe the sink
a1.sinks.k1.type = logger
# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100
# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1
3、创建文件夹 /home/hadoop/flume_test
4、启动，
命令：flume-ng agent -c ../conf -f spooldir-logger.conf -n a1 -Dflume.root.logger=INFO,console
5、在flume_test文件夹下新建文件
echo "11111" >> hello.txt
flume_test文件夹下多了个文件：hello.txt.COMPLETED
但flume报错了。
2017-03-22 21:55:50,005 (pool-3-thread-1) [ERROR - org.apache.flume.source.SpoolDirectorySource$SpoolDirectoryRunnable.run(SpoolDirectorySource.java:280)] FATAL: Spool Directory source r1: { spoolDir: /home/hadoop/flume_test }: Uncaught exception in SpoolDirectorySource thread. Restart or reconfigure Flume to continue processing.
java.lang.IllegalStateException:
File name has been re-used with different files. Spooling assumptions violated for /home/hadoop/flume_test/hello.txt.COMPLETED
跟踪抛出异常的源码，SpoolDirectorySource会启动一个线程轮询监控目录下的目标文件，当读取完该文件(readEvents)之后会对该文件进行重名(rollCurrentFile)，当重命名失败时会抛出IllegalStateException，被SpoolDirectoryRunnable catch重新抛出RuntimeException，导致当前线程退出，从源码看SpoolDirectoryRunnable是单线程执行的，因此线程结束后，监控目录下其他文件不再被处理。所以，再新建个 word.txt 文件，flume没有监听动作了。如下图，word.txt没有被重命名 word.txt.COMPLETED
正确的做法：
不要在flume_test文件夹下直接新建文件，写内容。在其他文件下新建，写好内容，mv 到flume_test文件夹下。
[hadoop@nbdo3 ~]$ cd testdata/
[hadoop@nbdo3 testdata]$ ll
total 4
-rw-rw-r--. 1 hadoop hadoop 71 Mar 10 20:19 hello.txt
[hadoop@nbdo3 testdata]$ cp hello.txt ../flume_test/
[hadoop@nbdo3 testdata]$ echo "123456778" >> world.txt
[hadoop@nbdo3 testdata]$ cp world.txt ../flume_test/
[hadoop@nbdo3 testdata]$
-------------
更多的Java，Angular，Android，大数据，J2EE，Python，数据库，Linux，Java架构师，：
http://www.cnblogs.com/zengmiaogen/p/7083694.html