Introducing Uno and Muchos

Author:   Mike Walch
Date:   21 Apr 2017

While Accumulo’s installation instructions are simple, it can be time consuming to install Accumulo given its requirement on Hadoop and Zookeeper being installed and running. For a one-time production installation, this set up time (which can take up to an hour) is not much of an inconvenience. However, it can become a burden for developers who need to frequently set up Accumulo to test code changes, switch between different versions, or start a fresh instance on a laptop.

Uno and Muchos are tools that ease the burden on developers of installing Accumulo and its dependencies. The names of Uno and Muchos indicate their use case. Uno is designed for running Accumulo on a single node while Muchos is designed for running Accumulo on a cluster. While Uno and Muchos will install by default the most recent stable release of Accumulo, Hadoop, and Zookeeper, it is easy to configure different versions to match a production cluster.

The sections below show how to use these tools. For more complete documentation, view their respective GitHub pages.

Uno

Uno is a command line tool that sets up Accumulo on a single machine. It can be installed by cloning the Uno git repo.

git clone https://github.com/apache/fluo-uno.git
cd fluo-uno

Uno works out of the box but it can be customized by modifying conf/uno.conf.

First, download the Accumulo, Hadoop, and Zookeeper tarballs from Apache by using the command below:

./bin/uno fetch accumulo

The fetch command places all tarballs in the downloads/ directory. Uno can be configured (in conf/uno.conf) to build an Accumulo tarball from a local git repo when fetch is called.

After downloading tarballs, the command below sets up Accumulo, Hadoop & Zookeeper in the install/ directory.

./bin/uno setup accumulo

Accumulo, Hadoop, & Zookeeper are now ready to use. You can view the Accumulo monitor at http://localhost:9995. You can configure your shell using the command below:

eval "$(./bin/uno env)"

Run uno stop accumulo to cleanly stop your cluster and uno start accumulo to start it again.

If you need a fresh cluster, you can run uno setup accumulo again. To kill your cluster, run uno kill.

Muchos

Muchos is a command line tool that launches an AWS EC2 cluster with Accumulo set up on it. It is installed by cloning its git repo.

git clone https://github.com/apache/fluo-muchos.git
cd fluo-muchos

Before using Muchos, create muchos.props in conf/ and edit it for your AWS environment.

cp conf/muchos.props.example conf/muchos.props
vim conf/muchos.props

Next, run the command below to launch a cluster in AWS.

muchos launch -c mycluster

After launching the cluster, set up Accumulo on it using the following command.

muchos setup

Use muchos ssh to ssh to the cluster and muchos terminate to terminate all EC2 nodes when you are finished.

Conclusion

Uno and Muchos automate installing Accumulo for development and testing. While not recommended for production use at this time, Muchos is a great reference for running Accumulo in production. System administrators can reference the Ansible code in Muchos to automate management of their own clusters.

View all posts in the news archive