Pan Yuxuan created FLUME-3054:
---------------------------------
Summary: hflushOrSync method in HDFS Sink should not treat
ClosedChannelException as an error
Key: FLUME-3054
URL: https://issues.apache.org/jira/browse/FLUME-3054
Project: Flume
Issue Type: Bug
Components: Sinks+Sources
Reporter: Pan Yuxuan
Fix For: v1.8.0
When use HDFS Sink in multiple threads, we face the error in log as below:
{code}
09 Feb 2017 13:44:14,721 ERROR [hdfs-hsProt6-call-runner-4]
(org.apache.flume.sink.hdfs.AbstractHDFSWriter.hflushOrSync:267) - Error while
trying to hflushOrSync!
09 Feb 2017 14:54:48,271 ERROR [SinkRunner-PollingRunner-DefaultSinkProcessor]
(org.apache.flume.sink.hdfs.AbstractHDFSWriter.isUnderReplicated:98) -
Unexpected error while checking replication factor
java.lang.reflect.InvocationTargetException
at sun.reflect.GeneratedMethodAccessor4.invoke(Unknown Source)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at
org.apache.flume.sink.hdfs.AbstractHDFSWriter.getNumCurrentReplicas(AbstractHDFSWriter.java:165)
at
org.apache.flume.sink.hdfs.AbstractHDFSWriter.isUnderReplicated(AbstractHDFSWriter.java:84)
at
org.apache.flume.sink.hdfs.BucketWriter.shouldRotate(BucketWriter.java:583)
at org.apache.flume.sink.hdfs.BucketWriter.append(BucketWriter.java:518)
at
org.apache.flume.sink.hdfs.HDFSEventSink.process(HDFSEventSink.java:418)
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:745)
Caused by: java.nio.channels.ClosedChannelException
at
org.apache.hadoop.hdfs.DFSOutputStream.checkClosed(DFSOutputStream.java:1665)
at
org.apache.hadoop.hdfs.DFSOutputStream.getCurrentBlockReplication(DFSOutputStream.java:2151)
at
org.apache.hadoop.hdfs.DFSOutputStream.getNumCurrentReplicas(DFSOutputStream.java:2140)
... 11 more
09 Feb 2017 14:54:48,277 ERROR [SinkRunner-PollingRunner-DefaultSinkProcessor]
(org.apache.flume.sink.hdfs.HDFSEventSink.process:459) - process failed
org.apache.flume.auth.SecurityException: Privileged action failed
at org.apache.flume.auth.UGIExecutor.execute(UGIExecutor.java:49)
at
org.apache.flume.auth.KerberosAuthenticator.execute(KerberosAuthenticator.java:63)
at org.apache.flume.sink.hdfs.BucketWriter$9.call(BucketWriter.java:676)
at java.util.concurrent.FutureTask.run(FutureTask.java:262)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)
Caused by: java.nio.channels.ClosedChannelException
at
org.apache.hadoop.hdfs.DFSOutputStream.checkClosed(DFSOutputStream.java:1665)
at org.apache.hadoop.fs.FSOutputSummer.write(FSOutputSummer.java:104)
at
org.apache.hadoop.fs.FSDataOutputStream$PositionCache.write(FSDataOutputStream.java:58)
at java.io.DataOutputStream.write(DataOutputStream.java:107)
at java.io.FilterOutputStream.write(FilterOutputStream.java:97)
at
org.apache.flume.serialization.BodyTextEventSerializer.write(BodyTextEventSerializer.java:71)
at
org.apache.flume.sink.hdfs.HDFSDataStream.append(HDFSDataStream.java:124)
at org.apache.flume.sink.hdfs.BucketWriter$7.call(BucketWriter.java:550)
at org.apache.flume.sink.hdfs.BucketWriter$7.call(BucketWriter.java:547)
at
org.apache.flume.sink.hdfs.BucketWriter$9$1.run(BucketWriter.java:679)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:415)
at
org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1858)
at org.apache.flume.auth.UGIExecutor.execute(UGIExecutor.java:47)
... 6 more
09 Feb 2017 14:54:48,280 ERROR [SinkRunner-PollingRunner-DefaultSinkProcessor]
(org.apache.flume.SinkRunner$PollingRunner.run:160) - Unable to deliver event.
Exception follows.
org.apache.flume.EventDeliveryException:
org.apache.flume.auth.SecurityException: Privileged action failed
at
org.apache.flume.sink.hdfs.HDFSEventSink.process(HDFSEventSink.java:463)
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:745)
Caused by: org.apache.flume.auth.SecurityException: Privileged action failed
at org.apache.flume.auth.UGIExecutor.execute(UGIExecutor.java:49)
at
org.apache.flume.auth.KerberosAuthenticator.execute(KerberosAuthenticator.java:63)
at org.apache.flume.sink.hdfs.BucketWriter$9.call(BucketWriter.java:676)
at java.util.concurrent.FutureTask.run(FutureTask.java:262)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
... 1 more
Caused by: java.nio.channels.ClosedChannelException
at
org.apache.hadoop.hdfs.DFSOutputStream.checkClosed(DFSOutputStream.java:1665)
at org.apache.hadoop.fs.FSOutputSummer.write(FSOutputSummer.java:104)
at
org.apache.hadoop.fs.FSDataOutputStream$PositionCache.write(FSDataOutputStream.java:58)
at java.io.DataOutputStream.write(DataOutputStream.java:107)
at java.io.FilterOutputStream.write(FilterOutputStream.java:97)
at
org.apache.flume.serialization.BodyTextEventSerializer.write(BodyTextEventSerializer.java:71)
at
org.apache.flume.sink.hdfs.HDFSDataStream.append(HDFSDataStream.java:124)
at org.apache.flume.sink.hdfs.BucketWriter$7.call(BucketWriter.java:550)
at org.apache.flume.sink.hdfs.BucketWriter$7.call(BucketWriter.java:547)
at
org.apache.flume.sink.hdfs.BucketWriter$9$1.run(BucketWriter.java:679)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:415)
at
org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1858)
at org.apache.flume.auth.UGIExecutor.execute(UGIExecutor.java:47)
... 6 more
{code}
The error shows that when call hflushOrSync(), the DataOutputStream was closed
by other threads, so DataOutputStream throws ClosedChannelException. But the
hdfs sink was not affected by the exception, DataOutputStream just throws the
ClosedChannelException as a remind. So we should ignore the
ClosedChannelException instead by logging as an error.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)