HPC Servers and Clusters with Bright Cluster Manager
Koi Computers’ engineering team excels at integrating leading cluster management software including Bright Cluster Manager (BCM). BCM automates the process of building and managing modern high-performance Linux clusters, eliminating complexity and enabling flexibility by allowing users to deploy complete Linux clusters over bare metal and manage them reliably. BCM combines provisioning, monitoring, and management capabilities in a single tool that spans the entire lifecycle of a Linux cluster. Highlights include:
Automated Cluster Provisioning & Setup
- Provisions nodes from bare metal
- Sets-up network, user directory & security
- Deploys workload managers & frameworks
- Runs pre-checks across the entire cluster prior to deployment
Automated Problem Detection And Isolation
- Provides detailed monitoring metrics & health-checks
- Reports resource usage by job and user
- Allows automated corrective actions to be taken
- Automatically detects and images replacement nodes
Automated Change Management
- 1-step propagation of updates
- Auto-update nodes without reimaging
- Easily add more servers & new components
- Replace failing nodes without disruption
- Prevents image-drift amongst nodes
- Supports mixed processor, accelerator and Linux combinations within the same cluster
- Run jobs on bare-metal, VMs or containers in the same cluster
- Host HPC, machine learning and data analytics applications on the same cluster
- Extend to the public cloud for additional resources
- Provision and manage edge compute as part of your central cluster
- Dynamically allocate resources where they’re needed based on demand
Bright Cluster Manager 9.1
Bright Computing introduced the latest iteration of BCM–Version 9.1, which further simplifies building and managing clusters. Key improvements include:
- The ability to automatically increase or decrease the number of nodes (servers) available to an HPC workload manager or to Kubernetes in a cluster. The allocation of nodes can be determined by demand and by policy.
- Enhancements that optimize performance, provide more granular control, and improve ease of use:
- Faster reallocation and re-purposing of nodes that share a common software image
- Ability to control and prioritize which cluster resources can be used by different jobs
- Ability to leverage priorities and job resource requirements from workload managers
- Easier setup of auto scaling via a wizard
- Eliminates silos in high-performance infrastructure, increases and optimizes overall compute utilization, and improves and accelerates responsiveness to end users’ needs.
Redfish Automated Bios Management
- Support for Redfish BIOS management that allows Bright administrators to view and update BIOS parameters on any/all nodes of the cluster that support the Redfish specification, from the command line or from the BrightView GUI.
- Reduces the complexity and the effort in updating BIOS settings across the cluster
Bright View UI Redesign
- The Bright View graphical admin interface has been redesigned to improve navigation and usability, making it more efficient to use.
- In BCM 9.1, the cluster monitoring function for the entire system can be offloaded from the system’s head node to a set of dedicated servers that perform system monitoring exclusively. This frees the head node to perform other duties and allows the system to continue scaling. If a dedicated monitoring node fails, the remaining monitoring nodes will take over monitoring of the orphaned compute nodes until the failed monitoring node is reinstated.
- Dedicated monitoring nodes automatically back-up monitoring data from the system, ensuring that monitoring data isn’t lost if the system goes down.
For more information on working with Koi Computers for servers and clusters integrated with cluster management software, including Bright Cluster Manager, contact us at 888-LOVE-KOI (888-568-3564); email: email@example.com or fill out the contact form at: https://koicomputers.com/contact-us/
Read the Bright Cluster Manager 9.1 Solution Brief.