Ryba boostraps and manages a full secured Hadoop cluster with one command. This is an Open-source software (OSS) project released under the new BSD license originally developed for one of the World largest utility company. It is used every day to manager and keep to date the cluster for every components.
Install Ryba locally or on a remote server and you are ready to go. It uses SSH to connect to each server of your cluster and it will install and check all the components you wish. You don't need to prepare your cluster nodes as long as a minimal installation of RHEL or CentOS is installed with a root user or a user with sudo access.
- Use secured comminication with SSH
- No database used, full distribution across multiple servers relying on GIT
- No agent or pre-installation required on your cluster nodes
- Version control all your configuration and modifications with GIT and NPM, the Node.js Package Manager
- Command-based to integrate with your Business Continuity Plan (BCP) and existing scripts
- For developer, as simple as learning Node.js and not a new framework
- Self-documented code written in Literate CoffeeScript
- Idempotent and executable on a running cluster without any negative impact
- Bootstrap the nodes from a fresh install
- Configure proxy environment if needed
- Optionnaly create a bind server (useful in Vagrant development environment)
- Install OpenLDAP and Kerberos and/or integrate with your existing infrastructure
- Deploy the latest Hortonworks Data Platform (HDP)
- Setup High Availabity for HDFS
- Integrate Kerberos with cross realm support
- Set IPTables rules and startup scripts
- Check the running components
- Provide convenient utilities such as global start/stop/status commands, distributed shell execution, ...
First download Node.js. You might need to adjust the name of the Node.js archive depending on the version you choose to install. Also, replace the path "/usr/local/node" to another location (eg "~/node") if you don't have the permission to write inside "/usr/local".
# Download the Node.js package wget --no-check-certificate https://nodejs.org/download/release/v6.2.2/node-v6.2.2-linux-x64.tar.gz # Extract the Node.js package tar xzf node-v6.2.2-linux-x64.tar.gz # Move Node.js into its final destination sudo mv node-v6.2.2-linux-x64 /usr/local/node # Add path to Node.js binary echo 'export PATH=/usr/local/node/bin:$PATH' >> ~/.bashrc # Source the update profile . ~/.bashrc # Check if node is installed node -v # Clean up uploaded archive rm -rf node-v6.2.2-linux-x64.tar.gz
If you are behind a proxy, configure the Node.js Pakage Manager (NPM) with the commands:
npm config set proxy http://proxy.company.com:8080 npm config set https-proxy http://proxy.company.com:8080
npm install to download the project dependencies.
Authentication Ryba configures every components to work with Kerberos when possible. All the components listed above (except Elasticsearch, MongoDB, Nagios in community version) does support Kerberos.
Authorization Since Ryba does support Apache Ranger, you can manage easily the Access Control List from Ranger Admin. Indeed Apache Ranger provides support for ACL administration for the main Big Data components under the Apache project.
Encryption Ryba configures TLS/SSL encryption for every service. You can generate (see an example on https://github.com/ryba-io/ryba-env-metal) or provide your own certificate, and Ryba will upload the certificates on the nodes and configure the components.
At the end of the ryba installation, you have a full Kerberized cluster with SSL encryption enabled.
Ryba does configure every service with High Availibity, if the service supports it. It does the configuration according to the layout of the cluster. Just define where you want the service to be installed, and Ryba does every step like installing, starting and checking.
Ryba has a check command which run components, to verifiy that it is rightly configured and running. Check can be port binding verification (for example port 50470 for the Hadoop HDFS Namenode), or complete functional test like launching mapreduce jobs on YARN.