fix bug where the curator config name was changed in one place but not another; make some info msgs into debug msgs; fix zkworker serialization

make RTR idempotent to multiple run requests for same task, because higher level things in the indexing service require this behaviour

1) Fix HadoopDruidIndexerConfig to no longer replace ":" with "_" on the segmentOutputDir.  The segmentOutputDir is user-supplied so they should have the ability to just not set a bad directory.

Remove retry logic out of the Remote Task Runner and introduce Task resources

add the ability to unmap mmapped files

put all concurrency test back  into server manager test, introduce better concurrency for reference counting

Conflicts:
	indexing-service/src/main/java/com/metamx/druid/indexing/worker/executor/ExecutorNode.java

Add a timeout in LoadQueuePeon to fail an assign if it takes too long

Toggle between compressed and non compressed service discovery

Fix path used when storing segments to HDFS