We're using HBaseTestingUtility's mini-cluster for testing some client code.
It works great, except that we sometimes get ugly error messages after our
tests finish.
I guessed this was because we weren't shutting down the mini-cluster, so I
added that in. But calling shutdownMiniCluster() still results in some
warnings and stack traces (although two of them seem to be from shutdown of the
HDFS cluster).
Is this normal/expected? Am I likely to waste more time than it is worth
trying to fix it? This is on 0.90.0.
Thanks.
joe
Sample:
17:31:51,213 WARN regionserver.handler.CloseRegionHandler:104 - Received CLOSE
for region Folder,,1300494707186.7b024fb711a7a7e962e4949ee17a05d3. but
currently not serving
17:31:51,288 WARN hbase.regionserver.HRegion:494 - Region
UserTable,,1300494705653.40e766ad609e0f46871d7a6b488fd669. already closed
17:31:51,290 WARN regionserver.handler.CloseRegionHandler:122 - Can't close
region: was already closed during close():
UserTable,,1300494705653.40e766ad609e0f46871d7a6b488fd669.
Shutting down the Mini HDFS Cluster
Shutting down DataNode 0
17:31:52,872 WARN server.datanode.DataNode:137 -
DatanodeRegistration(127.0.0.1:59036,
storageID=DS-541927814-10.132.173.136-59036-1300494697003, infoPort=59037,
ipcPort=59038):DataXceiveServer: java.nio.channels.AsynchronousCloseException
at
java.nio.channels.spi.AbstractInterruptibleChannel.end(AbstractInterruptibleChannel.java:185)
at
sun.nio.ch.ServerSocketChannelImpl.accept(ServerSocketChannelImpl.java:159)
at sun.nio.ch.ServerSocketAdaptor.accept(ServerSocketAdaptor.java:84)
at
org.apache.hadoop.hdfs.server.datanode.DataXceiverServer.run(DataXceiverServer.java:130)
at java.lang.Thread.run(Thread.java:680)
17:31:53,993 WARN server.namenode.FSNamesystem:2453 - ReplicationMonitor
thread received InterruptedException.java.lang.InterruptedException: sleep
interrupted
