flume-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexander Alten-Lorenz <wget.n...@gmail.com>
Subject Re: Getting "Checking file:conf/flume.conf for changes" message in loop
Date Thu, 02 May 2013 09:54:54 GMT
You may look into interceptors:
http://flume.apache.org/releases/content/1.3.0/apidocs/org/apache/flume/interceptor/Interceptor.html

Regards,
 Alex

On May 1, 2013, at 5:00 PM, Vikas Kanth <kanth_vikas@yahoo.co.in> wrote:

> Hi Jeff,
> 
> Thanks for the reply. Your suggestion worked.
> 
> I've got one more question. The logs generated at the HDFS, using exec/spooling,  are
in the following format :
> -rw-r--r--   3 vikas         66 2013-05-01 05:43 /tmp/FlumeData.1367412193479
> -rw-r--r--   3 vikas         67 2013-05-01 05:43 /tmp/FlumeData.1367412193480
> -rw-r--r--   3 vikas         61 2013-05-01 05:43 /tmp/FlumeData.1367412193481
> -rw-r--r--   3 vikas         44 2013-05-01 05:43 /tmp/FlumeData.1367412193482
> 
> Is there any way where I can copy the files (eg. com/org/test/flm/Sample.txt) from source
to HDFS in the same folder structure/size/name?
> If not, what is the alternative?
> 
> Thanks,
> Vikas
> From: Jeff Lord <jlord@cloudera.com>
> To: user@flume.apache.org; Vikas Kanth <kanth_vikas@yahoo.co.in> 
> Sent: Wednesday, 1 May 2013 7:05 AM
> Subject: Re: Getting "Checking file:conf/flume.conf for changes" message in loop
> 
> Vikas,
> 
> This message is normal and harmless.
> 
> 2013-04-29 08:26:11,868 (conf-file-poller-0) [DEBUG - org.apache.flume.conf.file.AbstractFileConfigurationProvider$FileWatcherRunnable.run(AbstractFileConfigurationProvider.java:188)]
Checking file:conf/flume.conf for changes
> 
> If you change your log settings to INFO level it will also not show up.
> 
> Regarding the reason for which you do not see the contents of your file in hdfs.
> One thing with the exec source and tail is that the events are buffered until 20 events
have been written to the cache. One way to work around this is to change the default from
20 -> 1
> 
> batchSize 	20	The max number of lines to read and send to the channel at a time
> 
> Alternatively there was a recent patch that sets the batchTimeout for exec source that
will let you flush the cache based on elapsed time. That fix is available on the latest version
of trunk.
> 
> -Jeff
> 
> 
> 
> On Mon, Apr 29, 2013 at 8:31 AM, Vikas Kanth <kanth_vikas@yahoo.co.in> wrote:
> Hi,
> 
> I am getting following message in loop. The source file hasn't moved to the destination.
> 
> 2013-04-29 08:24:41,346 (lifecycleSupervisor-1-0) [INFO - org.apache.flume.instrumentation.MonitoredCounterGroup.start(MonitoredCounterGroup.java:73)]
Component type: CHANNEL, name: Channel-2 started
> 2013-04-29 08:24:41,846 (conf-file-poller-0) [INFO - org.apache.flume.node.nodemanager.DefaultLogicalNodeManager.startAllComponents(DefaultLogicalNodeManager.java:141)]
Starting Sink HDFS
> 2013-04-29 08:24:41,847 (conf-file-poller-0) [INFO - org.apache.flume.node.nodemanager.DefaultLogicalNodeManager.startAllComponents(DefaultLogicalNodeManager.java:152)]
Starting Source tail
> 2013-04-29 08:24:41,847 (lifecycleSupervisor-1-3) [INFO - org.apache.flume.source.ExecSource.start(ExecSource.java:155)]
Exec source starting with command:tail -F /home/vkanth/temp/Sample2.txt
> 2013-04-29 08:24:41,850 (lifecycleSupervisor-1-3) [DEBUG - org.apache.flume.source.ExecSource.start(ExecSource.java:173)]
Exec source started
> 2013-04-29 08:24:41,850 (lifecycleSupervisor-1-0) [INFO - org.apache.flume.instrumentation.MonitoredCounterGroup.register(MonitoredCounterGroup.java:89)]
Monitoried counter group for type: SINK, name: HDFS, registered successfully.
> 2013-04-29 08:24:41,851 (lifecycleSupervisor-1-0) [INFO - org.apache.flume.instrumentation.MonitoredCounterGroup.start(MonitoredCounterGroup.java:73)]
Component type: SINK, name: HDFS started
> 2013-04-29 08:24:41,852 (SinkRunner-PollingRunner-DefaultSinkProcessor) [DEBUG - org.apache.flume.SinkRunner$PollingRunner.run(SinkRunner.java:143)]
Polling sink runner starting
> 2013-04-29 08:25:11,855 (conf-file-poller-0) [DEBUG - org.apache.flume.conf.file.AbstractFileConfigurationProvider$FileWatcherRunnable.run(AbstractFileConfigurationProvider.java:188)]
Checking file:conf/flume.conf for changes
> 2013-04-29 08:25:41,861 (conf-file-poller-0) [DEBUG - org.apache.flume.conf.file.AbstractFileConfigurationProvider$FileWatcherRunnable.run(AbstractFileConfigurationProvider.java:188)]
Checking file:conf/flume.conf for changes
> 2013-04-29 08:26:11,868 (conf-file-poller-0) [DEBUG - org.apache.flume.conf.file.AbstractFileConfigurationProvider$FileWatcherRunnable.run(AbstractFileConfigurationProvider.java:188)]
Checking file:conf/flume.conf for changes
> .......
> .......
> 
> 
> Flume.conf:
> agent1.sources = tail
> agent1.channels = Channel-2
> agent1.sinks = HDFS
> 
> agent1.sources.tail.type = exec
> agent1.sources.tail.command = tail -F /home/vikas/temp/Sample2.txt
> agent1.sources.tail.channels = Channel-2
> 
> agent1.sinks.HDFS.channel = Channel-2
> agent1.sinks.HDFS.type = hdfs
> agent1.sinks.HDFS.hdfs.path = hdfs://dev-pub01.xyz.abc.com:8020/tmp
> agent1.sinks.HDFS.hdfs.file.fileType = DataStream
> 
> agent1.channels.Channel-2.type = memory
> agent1.channels.Channel-2.capacity = 1000
> agent1.channels.Channel-2.transactionCapacity=10
> 
> Command:
> bin/flume-ng agent --conf ./conf/ -f conf/flume.conf -Dflume.root.logger=DEBUG,console
-n agent1
> 
> Please let me know if I am missing something.
> 
> Thanks,
> Vikas
> 
> 
> 
> 

--
Alexander Alten-Lorenz
http://mapredit.blogspot.com
German Hadoop LinkedIn Group: http://goo.gl/N8pCF


Mime
View raw message