HugeTable

Open-source implementation of Google's BigTable.

Getting Started

Obtaining the source code

Download the source code by running the following code in your command prompt:

git clone https://github.com/fabianishere/in4391.git

or simply grab a copy of the source code as a Zip file.

Building the project

For building the project, we use Gradle. To perform a build, enter the following in your command prompt in the root directory of the project:

./gradlew build

To build the binaries, make sure you also run the following command:

./gradlew installDist

Setup the environment

HugeTable requires the presence of a Hadoop HDFS cluster for distributed file storage and a ZooKeeper cluster for synchronization between the nodes. To simplify the deployment of HugeTable within a test environment, you may use the single node test environment we provide in the repository.

To start the test environment, please run the following command:

build/install/hugetable/bin/htable-test-env    

which will start a HDFS node on port 9000 and a ZooKeeper node on port 2181.

Start a HugeTable cluster

The next step is to start one or multiple HugeTable nodes, which together will form a cluster. You may start a node as follows:

build/install/hugetable/bin/htable-server-cli --zookeeper localhost:2181 --hadoop hdfs://localhost:9000 --port 8080

Make sure you specify the correct ports and don't use duplicate ports for different nodes.

Communicating with a HugeTable cluster

You may communicate with the HugeTable cluster by using the command-line client (htable-client-cli) or the Scala client library (htable-client). To use the command-line interface interactively, enter the following command in your command prompt:

build/install/hugetable/bin/htable-client-cli -i --zookeeper localhost:2181   

Make sure again that you specify the correct ZooKeeper port. Then, enter help to obtain the possible commands you can execute in the cluster.

License

HugeTable is available under the MIT license.