Bright Cluster Manager and Bright Cluster Manager for Data Science contain tools and applications to facilitate the installation, administration, and monitoring of a cluster. Also, Bright Cluster Manager provides users with an optimal environment for developing and running applications that require extensive computational resources.
Bright Cluster Manager for Data Science lets you administer data science clusters as a single entity, provisioning the hardware, operating system, HPC, data analytics, and deep learning software from a single interface. This capability makes it easier to build a reliable, clustered data science infrastructure. When your system is up and running, the Bright cluster management daemon monitors every aspect of every node. The daemon reports any problems that it detects in the software or the hardware so that you can act and keep your infrastructure healthy.
To prepare systems for software deployment using Bright software, ensure that each server is racked, powered, and networked so that the server can download software packages from the Internet or from a full Red Hat Enterprise Linux mirror.
Also, Bright View GUI users can manage hardware, Slurm, and Kubernetes, all from a dashboard, simplifying management and operations, as shown in the following figure:
Figure 3. Bright View GUI for Slurm job scheduling and cluster management
Other features of Bright Cluster Manager and Bright Cluster Manager for Data Science include:
The Cluster Management Shell (CMSH) provides the movenodes command to move a node:
movenodes - Move nodes from one overlay to another
movenodes overlayA overlayB node001 Move node001 from overlayA to overlayB
movenodes overlayA overlayB -n node00[1-3] Move node001..node003 from overlayA to overlayB
Switching a node from Slurm to Kubernetes and back takes approximately 5 to 10 seconds.
For more information, see Documentation on the Bright Computing website.