flume1.6.0-agent多channel 多sink问题

u010237107 2015-12-16 10:54:46
需求是这样的,客户端agent到collect agent,然后collect agent分两路,一路到hdfs 供离线分析,一路经kafka消息中间件到storm实时分析。

现在遇到的情况是,collect agent 单独得写配置到hdfs,或者到kafka 都是没有任何问题的,但唯独合在一起(1个sources,2个channel,2个sink)却出现问题了,查看flume的logs没发现有价值的东西,我认识卡在某个步骤了,非常得匪夷所思啊。

下面先贴上flume logs文件的内容

16 十二月 2015 10:45:57,414 INFO  [lifecycleSupervisor-1-0] (org.apache.flume.node.PollingPropertiesFileConfigurationProvider.start:61)  - Configuration provider starting
16 十二月 2015 10:45:57,427 INFO [conf-file-poller-0] (org.apache.flume.node.PollingPropertiesFileConfigurationProvider$FileWatcherRunnable.run:133) - Reloading configuration file:conf/myconf/flume_colletc_test.conf
16 十二月 2015 10:45:57,435 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k2
16 十二月 2015 10:45:57,436 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,436 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,436 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,436 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,437 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k2
16 十二月 2015 10:45:57,437 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,437 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,437 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k2
16 十二月 2015 10:45:57,437 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k2
16 十二月 2015 10:45:57,437 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,438 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,438 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,438 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,440 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k2
16 十二月 2015 10:45:57,440 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k2
16 十二月 2015 10:45:57,441 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,441 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:1017) - Processing:k1
16 十二月 2015 10:45:57,441 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.addProperty:931) - Added sinks: k2,k1 Agent: collectorMainAgent
16 十二月 2015 10:45:57,443 WARN [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration$AgentConfiguration.isValid:337) - Agent configuration for 'collectorMainAgent' does not contain any valid channels. Marking it as invalid.
16 十二月 2015 10:45:57,443 WARN [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration.validateConfiguration:128) - Agent configuration invalid for agent 'collectorMainAgent'. It will be removed.
16 十二月 2015 10:45:57,445 INFO [conf-file-poller-0] (org.apache.flume.conf.FlumeConfiguration.validateConfiguration:141) - Post-validation flume configuration contains configuration for agents: []
16 十二月 2015 10:45:57,445 WARN [conf-file-poller-0] (org.apache.flume.node.AbstractConfigurationProvider.getConfiguration:133) - No configuration found for this host:collectorMainAgent
16 十二月 2015 10:45:57,453 INFO [conf-file-poller-0] (org.apache.flume.node.Application.startAllComponents:138) - Starting new configuration:{ sourceRunners:{} sinkRunners:{} channels:{} }


然后贴上我 collect的配置文件
# collectorMainAgent
collectorMainAgent.channels = c1,c2
collectorMainAgent.sources = s2
collectorMainAgent.sinks =k2,k1

# collectorMainAgent AvroSource
collectorMainAgent.sources.s2.type = avro
collectorMainAgent.sources.s2.bind = 172.21.99.134
collectorMainAgent.sources.s2.port = 44444
collectorMainAgent.sources.s2.channels = c1,c2


# defined for kafka channel by Cesar.X 2015/12/16
collectorMainAgent.channels.c1.type = memory
collectorMainAgent.channels.c1.capacity = 1000
collectorMainAgent.channels.c1.transactionCapacity = 100

#定义flume-kafka by Cesar.X 2015/12/15
collectorMainAgent.sinks.k2.type = org.apache.flume.sink.kafka.KafkaSink
collectorMainAgent.sinks.k2.topic = test-topic
collectorMainAgent.sinks.k2.brokerList = master:9092,slave1:9092,slave2:9092
collectorMainAgent.sinks.k2.requiredAcks = 1
collectorMainAgent.sinks.k2.batchSize = 20
collectorMainAgent.sinks.k2.channel = c1


# collectorMainAgent FileChannel
#defined for hdfs by Cesar.X 2015/12/16
collectorMainAgent.channels.c2.type = file
collectorMainAgent.channels.c2.checkpointDir =/usr/local/hadoop/apache-flume-1.6.0-bin/fchannel/spool/checkpoint
collectorMainAgent.channels.c2.dataDirs = /usr/local/hadoop/apache-flume-1.6.0-bin/fchannel/spool/data
#collectorMainAgent.channels.c2.capacity = 200000000
collectorMainAgent.channels.c2.capacity = 1000
collectorMainAgent.channels.c2.transactionCapacity=6000
collectorMainAgent.channels.c2.checkpointInterval=60000

collectorMainAgent.sinks.k1.type = hdfs
collectorMainAgent.sinks.k1.channel = c2
collectorMainAgent.sinks.k1.hdfs.path = hdfs://mycluster/user/flume/events/%Y-%m-%d
collectorMainAgent.sinks.k1.hdfs.filePrefix = events
#collectorMainAgent.sinks.k1.hdfs.round = true
#collectorMainAgent.sinks.k1.hdfs.roundValue = 10
#collectorMainAgent.sinks.k1.hdfs.roundUnit = minute
# Number of seconds to wait before rolling current file (0 = never roll based on time interval)
collectorMainAgent.sinks.k1.hdfs.rollInterval=600
# File size to trigger roll, in bytes (256Mb),对于hdfs来说,应该尽量避免小文件的产生
# 这边有很多触发器,依次满足,先满足hdfs 文件大小256MB分割
collectorMainAgent.sinks.k1.hdfs.rollSize = 268435456
# never roll based on number of events(0 = never roll based on number of events)
collectorMainAgent.sinks.k1.hdfs.rollCount = 0
# Timeout after which inactive files get closed (0 = disable automatic closing of idle files)
# 持续60分钟没动作分割
collectorMainAgent.sinks.k1.hdfs.idleTimeout = 3600
collectorMainAgent.sinks.k1.hdfs.writeFormat = Text
collectorMainAgent.sinks.k1.hdfs.inUseSuffix =.txt
collectorMainAgent.sinks.k1.hdfs.fileType = DataStream
#collectorMainAgent.sinks.HDFS.hdfs.batchSize = 1000
#必须有时间戳,否则会报错
collectorMainAgent.sinks.k1.hdfs.useLocalTimeStamp = true


分就这么多,希望大家谅解!!
...全文
1758 6 打赏 收藏 转发到动态 举报
写回复
用AI写文章
6 条回复
切换为时间正序
请发表友善的回复…
发表回复
a13575018718 2016-12-30
  • 打赏
  • 举报
回复
你这个是哪里改成空格啊 哪个地方逗号出错了啊
zhtwave 2016-03-29
  • 打赏
  • 举报
回复
_游侠 2016-03-15
  • 打赏
  • 举报
回复
该是空格。
  • 打赏
  • 举报
回复
其实也挺好的,算是一个分享。
夜无边CN 2015-12-16
  • 打赏
  • 举报
回复
引用 1 楼 u010237107 的回复:
2B了,大意了,是逗号的问题!!!!
u010237107 2015-12-16
  • 打赏
  • 举报
回复
2B了,大意了,是逗号的问题!!!!

20,811

社区成员

发帖
与我相关
我的任务
社区描述
Hadoop生态大数据交流社区,致力于有Hadoop,hive,Spark,Hbase,Flink,ClickHouse,Kafka,数据仓库,大数据集群运维技术分享和交流等。致力于收集优质的博客
社区管理员
  • 分布式计算/Hadoop社区
  • 涤生大数据
加入社区
  • 近7日
  • 近30日
  • 至今
社区公告
暂无公告

试试用AI创作助手写篇文章吧