MeT is a Cloud-enabled framework that can be used alone or in conjunction with OpenStack for the automatic and heterogeneous reconfiguration of HBase. MeT is an workload aware system that provides automatic elasticity for the HBase NoSQL database. MeT not only adds and removes nodes automatically and according to system load, but also reconfigures them according to the observed workloads. As a result, it achieves a significant increase in overall system performance.
The current version of MeT is ready to work with HBase and OpenStack.
MeT has a Java module to bridge MeT and HBase.
Such module has the following dependencies:
- commons-configuration-1.6.jar
- commons-lang-2.5.jar
- commons-logging-1.1.1.jar
- hadoop-core-1.0.1.jar
- hbase-0.92.0-cdh4b1.jar
- log4j-1.2.15.jar
- py4j0.7.jar
- slf4j-api-1.6.1.jar
- slf4j-log4j12-1.6.1.jar
- zookeeper-3.4.1-cdh4b1.jar
MeT is mainly implemented in Python.
The following dependencies must be installed:
- paramiko
- py4j
- novaclient.v1_1 (OpenStack Client)
Just compile the JMeTGlue.java class. No need for extra build work.
Folder 'config' holds a set of configuration files. Folder 'tmp' holds hbase configuration files that MeT uses and copies to the actual running machines. These should be the only places where you would need to tune MeT.
-
Run Java Gateway (JMeTGlue)
-
./runner.sh (Please check if 'runner.sh' is considering the right py4j and paramiko paths)
Updated source and an issue tracker are available at:
https://github.com/fmaia/MeT
Your feedback is welcome.
#Contact
Francisco Cruz ([email protected])
Francisco Maia ([email protected])
Miguel Matos ([email protected])
Ricardo Vilaça ([email protected])
Joao Paulo ([email protected])