Source code for the internal HBase offline compacter.  This runs
during off-peak hours and cycles through all the region servers and
kicks off region compactions, one region at a time.  It then pauses
for a few seconds (configurable) and repeats the process till all the
regions are done.  The program also queries the current compaction
queue on the servers (via jmx) and skips the server if the queue size
is too high (configurable).  Also, it only runs during off-peak hours
(configurable as well).

The program expects JMX credentials in a text file with the format

role password

The name of this text file is hard coded in the JMXQuery.java file.

You will probably need the following libraries to get this to run.

JSAP-2.1.jar
commons-collections-3.2.1.jar
commons-lang-2.6.jar
commons-logging-1.1.1.jar
hadoop-core-0.20-append-r1056497.jar
hbase-0.90.3.jar
log4j-over-slf4j-1.6.1.jar
logback-classic-0.9.24.jar
logback-core-0.9.24.jar
slf4j-api-1.6.1.jar
zookeeper-3.3.1.jar