For the majority of our customers, the usual way of installing a Bright cluster is by downloading the Bright ISO, going through a head node installation, and then PXE booting your compute nodes to allow them to join the cluster.
For clusters in the cloud, this is not possible because the head node needs to be initialized using a pre-cooked image, and nodes don't PXE boot.
One solution to this problem is to bootstrap a Bright cluster, using the Bright Customer Portal and fill in the form under Cluster on Demand. If you are uncomfortable putting in details such as your AWS or Microsoft Azure authentication information in the Bright website – and rest assured that after we process the form we don't hold on to the details - there is another solution: Use the cluster on demand (COD) Docker container image.
The COD Docker image is available from Dockerhub and can be used on any Linux machine which has Docker installed. The image contains the COD utility and all of its dependencies, which allows users to quickly spin up any number of Bright clusters with various configurations in AWS, Azure, or OpenStack (and in the future also in VMWare).
The utility interacts with the AWS, OpenStack, or Azure API to create the cluster. It takes care of not only creating the instances but also (for example) the networks, the resource groups, etcetera. And there are many command line options for fine-tuning.
So, if you have the desire to quickly launch a Bright cluster in AWS or Azure, then this is the tool that you can use to bootstrap those clusters.
It’s easy to get started. You will require a Bright product key, and there are three options for this: Purchase one, request an evaluation product key, or request an Easy8 product key. In this last case, you can spin up 8 nodes free of charge. Visit the Bright Customer Portal to get started.
The power of this is exciting. For example, if you want to carry out a calculation and you have no physical infrastructure, you can quickly spin up a cluster of any size in AWS or Azure by using the COD Docker container.
To find out more about this innovation, please contact us.