The radanalytics.io projects are all focused on increasing your ability to design, deploy, and maintain insightful data-driven applications. The foundation of this effort are infrastructure projects which simplify the process of deploying and managing data processing frameworks.
This introduction will help you learn about the primary radanalytics.io projects, how to install them, and how to utilize them with your own applications.
Although the projects page contains all the collected efforts of the radanalytics.io community, there are several applications which deserve to be highlighted as they are key to beginning your journey here.
With the Oshinko family of applications you can create, scale, and destroy Apache Spark clusters. These clusters can then be used by your applications within an OpenShift project by providing a simple connection URL to the cluster. There are multiple paths to achieving this: browser based graphical interface, command line tool, and a RESTful server.
To begin your exploration, we recommend starting with the oshinko-webui application. The oshinko-webui is a self-contained deployment of the Oshinko technologies. An OpenShift user can deploy the oshinko-webui container into their project and then access the server with a web browser. Through the browser interface you will be able to manage Apache Spark clusters within your project.
Another important part of Oshinko to highlight is the oshinko-s2i repository and associated images which implement the source-to-image workflow for Apache Spark based applications. These images enable you to create full applications that can be built, deployed and upgraded directly from a source repository.
Download the oc command and login to an existing cluster with
oc login https://.... Alternatively, setup your own cluster with
oc cluster up or the
Minishift tool(Windows and Mac users
may find this option easiest to start).
First, install all the Oshinko resources into your project -
oc create -f https://radanalytics.io/resources.yaml
This creates the latest versions of the Oshinko S2I (source-to-image) templates and the Oshinko Web UI application, as well as a ServiceAccount and RoleBinding needed for creation and management of Apache Spark clusters. Follow this alternative to use components from a specific Oshinko release.
Second, start the Oshinko Web UI application -
oc new-app oshinko-webui
This creates the Oshinko Web UI, which can be used to manually create and manage Apache Spark clusters.
Oshinko Web UI
With Oshinko now installed and running in your OpenShift project, we recommend checking out the tutorial applications that the radanalytics.io community have created. These tutorials will you show how to deploy and utilize an insightful data-driven application with Oshinko and OpenShift.
You can use
resources.yaml to install Oshinko components from a specific release beginning with v0.2.6.
OSHINKO_VERSION to the release version, for example:
Then run this command to generate and process a modified resources.yaml:
wget https://radanalytics.io/resources.yaml -qO - \ | sed -r -e "s@(radanalyticsio/radanalytics.*spark)(:stable)@\1:$OSHINKO_VERSION@" \ | sed -r -e "s@(radanalyticsio/oshinko-webui)(:stable)@\1:$OSHINKO_VERSION@" \ | oc create -f -