本文共 1351 字，大约阅读时间需要 4 分钟。

原理

案例

hadoop01上的Flume-1监控文件/opt/module/group.log，

实现

# agent1# 定义sources、channels、sinksa1.sources = r1a1.channels = c1a1.sinks = k1# 配置sourcea1.sources.r1.type = TAILDIR# 日志索引记录a1.sources.r1.positionFile = /opt/module/flume/inode/taildir_position.json# 文件组a1.sources.r1.filegroups = f1a1.sources.r1.filegroups.f1 = /opt/module/group.log# 配置channela1.channels.c1.type = memory# 配置sinka1.sinks.k1.type = avroa1.sinks.k1.hostname = hadoop03a1.sinks.k1.port = 6666# 配置source、channel、sink关系a1.sources.r1.channels = c1a1.sinks.k1.channel = c1

# agent2# 定义sources、channels、sinksa1.sources = r1a1.channels = c1a1.sinks = k1# 配置sourcea1.sources.r1.type = netcata1.sources.r1.bind = hadoop02a1.sources.r1.port = 44444# channela1.channels.c1.type = memory# sinka1.sinks.k1.type = avroa1.sinks.k1.hostname = hadoop03a1.sinks.k1.port = 6666# 配置关系a1.sources.r1.channels = c1a1.sinks.k1.channel = c1

# agent3# 定义sources、channels、sinksa1.sources = r1a1.channels = c1a1.sinks = k1# 配置sourcea1.sources.r1.type = avroa1.sources.r1.bind = hadoop03a1.sources.r1.port = 6666# channela1.channels.c1.type = memory# sinka1.sinks.k1.type = logger# 配置关系a1.sources.r1.channels = c1a1.sinks.k1.channel = c1

先启动数据接收端，再启动数据发送端。

转载地址：https://blog.csdn.net/FlatTiger/article/details/114022601 如侵犯您的版权，请留言回复原文章的地址，我们会给您删除此文章，给您带来不便请您谅解！