FLUME-1630 (in flume 1.3.1) hopefully improved re-configuration, but in general I agree, restart is better.
On Wed, Jan 16, 2013 at 6:11 PM, Juhani Connolly <[email protected]> wrote: > Switching configuration on a running node is pretty buggy, I would recommend > just restarting flume, as while sometimes it will work there are issues like > components not getting properly shut down even when removed from the config. > > > On 01/17/2013 05:41 AM, Yatchmenoff, Sam wrote: > > I have Flume 1.2.0 running in a production system with 3 collectors fed by > ~30 agents running on our application servers. If I make a change to the > node configuration on the collectors, when the configuration is reloaded > automatically, the collectors will occasionally fail and repeatedly report > the following exception: > > 2013-01-16 20:31:22,353 ERROR flume.SinkRunner: Unable to deliver event. > Exception follows. > org.apache.flume.EventDeliveryException: Failed to process transaction > at org.apache.flume.sink.RollingFileSink.process(RollingFileSink.java:218) > at > org.apache.flume.sink.DefaultSinkProcessor.process(DefaultSinkProcessor.java:68) > at org.apache.flume.SinkRunner$PollingRunner.run(SinkRunner.java:147) > at java.lang.Thread.run(Thread.java:679) > Caused by: java.io.IOException: Stream Closed > at java.io.FileOutputStream.writeBytes(Native Method) > at java.io.FileOutputStream.write(FileOutputStream.java:297) > at java.io.BufferedOutputStream.flushBuffer(BufferedOutputStream.java:82) > at java.io.BufferedOutputStream.write(BufferedOutputStream.java:126) > at java.io.FilterOutputStream.write(FilterOutputStream.java:97) > at > org.apache.flume.serialization.BodyTextEventSerializer.write(BodyTextEventSerializer.java:71) > at org.apache.flume.sink.RollingFileSink.process(RollingFileSink.java:195) > ... 3 more > > After about a dozen of those, I will start seeing this exception: > > 2013-01-16 20:32:27,374 ERROR flume.SinkRunner: Unable to deliver event. > Exception follows. > org.apache.flume.EventDeliveryException: Unable to rotate file > /mnt/rawlog/1358365369665-49 while delivering event > at org.apache.flume.sink.RollingFileSink.process(RollingFileSink.java:155) > at > org.apache.flume.sink.DefaultSinkProcessor.process(DefaultSinkProcessor.java:68) > at org.apache.flume.SinkRunner$PollingRunner.run(SinkRunner.java:147) > at java.lang.Thread.run(Thread.java:679) > Caused by: java.io.IOException: Stream Closed > at java.io.FileOutputStream.writeBytes(Native Method) > at java.io.FileOutputStream.write(FileOutputStream.java:297) > at java.io.BufferedOutputStream.flushBuffer(BufferedOutputStream.java:82) > at java.io.BufferedOutputStream.flush(BufferedOutputStream.java:140) > at org.apache.flume.sink.RollingFileSink.process(RollingFileSink.java:149) > ... 3 more > > > Here is configuration for that agent: > > agent1.sources = source1 source2 > agent1.sinks = sink1 > agent1.channels = channel1 > > # Describe/configure source1 > agent1.sources.source1.type = avro > agent1.sources.source1.bind = 0.0.0.0 > agent1.sources.source1.port = 35853 > > agent1.sources.source2.type = netcat > agent1.sources.source2.bind = localhost > agent1.sources.source2.port = 35854 > agent1.sources.source2.max-line-length = 524288 > > # Describe sink1 > agent1.sinks.sink1.type = FILE_ROLL > agent1.sinks.sink1.sink.directory = /mnt/rawlog > agent1.sinks.sink1.sink.rollInterval = 60 > > # Use a channel which buffers events in memory > agent1.channels.channel1.type = file > agent1.channels.channel1.checkpointDir = > /mnt/flume-ng/file-channel1/checkpoint > agent1.channels.channel1.dataDirs = /mnt/flume-ng/file-channel1/data > agent1.channels.channel1.capacity = 100000 > > # Bind the source and sink to the channel > agent1.sources.source1.channels = channel1 > agent1.sources.source2.channels = channel1 > agent1.sinks.sink1.channel = channel1 > > Any ideas about what's causing this exception would be greatly appreciated. > > -- Apache MRUnit - Unit testing MapReduce - http://incubator.apache.org/mrunit/
