I have faced the same situation in the past. can you try to use flume V1.2 and see if you get the same exception.

On Mon, Mar 9, 2015 at 7:26 AM, Marina <ppine7@yahoo.com> wrote:
I have configured Flume to "tail -f" logs from my Varnish server - pretty much standard Apache HTTP logs.
However, sometimes Flume chokes on some special characters and dies - stops processing new log entries.

See below for a stack trace.

It seems like this exact issue was reported as Flume bug in 1.4.x version:
and it was marked as resolved in 1.5.0 version.
The version I am using is Flume 1.5.2 - and I am still seeing this issue...

Could somebody confirm/deny if what I am seeing is the same issue and should have been fixed? OR is this completely different?

Thank you!

06 Mar 2015 18:16:57,820 ERROR [pool-3-thread-1] (org.apache.flume.source.SpoolDirectorySource$SpoolDirectoryRunnable.run:256)  - FATAL: Spool Directory source r1: { spoolDir: /data1/varnish-logs-active }: Uncaught exception in SpoolDirectorySource thread. Restart or reconfigure Flume to continue processing.
java.nio.charset.MalformedInputException: Input length = 1
at java.nio.charset.CoderResult.throwException(CoderResult.java:260)
at org.apache.flume.serialization.ResettableFileInputStream.readChar(ResettableFileInputStream.java:195)
at org.apache.flume.serialization.LineDeserializer.readLine(LineDeserializer.java:134)
at org.apache.flume.serialization.LineDeserializer.readEvent(LineDeserializer.java:72)
at org.apache.flume.serialization.LineDeserializer.readEvents(LineDeserializer.java:91)
at org.apache.flume.client.avro.ReliableSpoolingFileEventReader.readEvents(ReliableSpoolingFileEventReader.java:238)
at org.apache.flume.source.SpoolDirectorySource$SpoolDirectoryRunnable.run(SpoolDirectorySource.java:227)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:439)
at java.util.concurrent.FutureTask$Sync.innerRunAndReset(FutureTask.java:317)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:150)