TrinityX Cluster Manager (Cluster Management System)
What’s TrinityX Cluster Manager?
TrinityX Cluster Manager is a tool developed by ClusterVision for building and managing HPC and AI systems. It streamlines configuration management and provisioning, provides real-time system monitoring, and offers an integrated management interface combined with a user portal significantly reducing operational costs.
Significantly Reduce Operational Costs
This software enables efficient deployment and management of clusters ranging from small systems with just a few nodes to large-scale environments with over 1,000 nodes. Even GPU-accelerated, large-scale cluster systems can be built and operated with ease, allowing you to greatly reduce operational costs.
Start Operations Immediately
TrinityX provides everything required for a Linux-based HPC or AI cluster environment in a single, turnkey solution. It natively supports OS management (RHEL and compatible distributions, Ubuntu), job schedulers (Slurm and PBS), CUDA and InfiniBand drivers and software stacks, user portals such as Open OnDemand, and parallel storage systems including BeeGFS, Lustre, and Spectrum Scale.
With these components seamlessly integrated, you can build and launch your system quickly and efficiently.
Key Features
Turn thousands of servers into a complete supercomputing environment with a single, integrated solution
- Ultra-fast provisioning capable of deploying large numbers of nodes simultaneously
- Pre-configured Slurm for maximizing cluster utilization
- Supports Singularity and Docker-based container environments
- Parallel file system support: BeeGFS, Lustre, IBM Spectrum Scale (GPFS)
- Comprehensive hardware integration including IPMI, InfiniBand, and GPUs
- HPC user environment: Open OnDemand, development tools, and MPI libraries (OpenMPI, Intel MPI, etc.)
- High availability for controller, storage, and login nodes
- Connects to existing authentication systems with minimal configuration
Comprehensive Cluster Management
- A fully integrated monitoring and metrics system for tracking critical events and resource usage
- Automatic discovery and unified management of nodes and switches
- Extensive GPU status monitoring, including load, temperature, and health checks
Implementation Consulting Services
Pacific Teck provides technical consulting services for TrinityX deployments, including initial system setup and configuration tailored to your specific use cases and operational requirements.
Feature Details
Seamless and Secure Cluster Deployment
TrinityX provides a user-friendly interface that guides you step-by-step through cluster setup and configuration. The resulting environment is suitable not only for traditional HPC workloads but also for modern AI applications. TrinityX also supports complex architectures out of the box, including multi-GPU configurations, InfiniBand/RDMA networks, and high-performance network switches, along with their monitoring and management.
For provisioning, TrinityX leverages the BitTorrent protocol, enabling rapid deployment of HPC/AI environments across clusters ranging from small systems with around 5 nodes to large-scale installations with more than 1,000 nodes—often completing in just a few minutes.
This deployment process is further secured through TPM-managed PXE boot, ensuring a trusted and secure initialization of all nodes.
User Portal
TrinityX includes the industry-standard Open OnDemand user portal, enabling seamless file input/output as well as job submission and management. Users can launch remote desktops via VNC, submit Jupyter Notebook sessions as Slurm jobs, and open a terminal to the controller node directly from their browser.
When accessed with an administrator account, the portal also provides entry to cluster management functions, including configuration management and system monitoring.
Centralized Cluster Monitoring and Reporting
TrinityX provides comprehensive monitoring of system health and performance across the entire cluster. The dashboard offers real-time visibility into cluster metrics, resource utilization, and workload distribution, and can be customized to suit your operational needs.
These insights help identify bottlenecks and optimize resource allocation for more efficient system operation.
Efficient Power and Resource Management
TrinityX provides the insights needed to manage cluster resources efficiently and maximize utilization. With features such as resource allocation, quota management, and automated resource provisioning, it helps minimize operational overhead while ensuring optimal system performance.
Advanced Scheduling and Load Balancing
TrinityX integrates and monitors Slurm and (Open)PBS Pro as standard cluster schedulers. It distributes workloads efficiently across the entire cluster and provides historical insights into queue activity and job accumulation.
These capabilities make it easy to analyze usage patterns and fine-tune scheduling configurations to match your operational needs.
Scalability and Open Source
TrinityX is a cluster manager built from a collection of open-source software components, offering the flexibility and scalability needed to meet your unique requirements. Its modular architecture makes it easy to customize and extend functionality, as well as integrate with other tools and systems.
Supported by an active open-source community, TrinityX continues to evolve, ensuring a reliable and future-proof solution.
Comprehensive Storage Support
TrinityX supports widely used parallel file systems in HPC, including BeeGFS, Lustre, and IBM Spectrum Scale (GPFS).
24/7 Managed Services
TrinityX provides a managed support framework that enables SLA-driven remote system management (PHASE III), delivering reliable operation around the clock, 365 days a year.
TrinityX Cluster Manager Support Levels
| TrinityX Commercial Support | TrinityX Community Edition | |
|---|---|---|
| Support | Enhanced Support | Self-Managed |
| Response Time | Next Business Day | – |
| Support & Discussion Channels | Email, Microsoft Teams | GitHub |
| TrinityX Code Support | 〇 | – |
| TrinityX Cluster Management Portal | 〇 | 〇 |
| ntegration of OpenHPC and Open OnDemand with TrinityX*1 | 〇 | 〇 |
| GPU, InfiniBand, and OpenHPC Tools: CUDA, ROCm, Slurm, MPI | 〇 | – |
| Health Check and Installation Support | 〇 | – |
| Diagnostics and Troubleshooting | 〇 | – |
| Parallel File System Installation, Tuning, and Support | Optional Services | – |
| End-User Application Integration, Tuning, and Testing | Optional Services | – |
| Data Backup Design, Integration, and Support | Optional Services | – |
| Comprehensive Administrator and User Training | Optional Services | – |
| Capacity Planning / User Application Benchmarking | Optional Services | – |
*1 If you require best-effort support for Open OnDemand under the TrinityX commercial support plan, we can provide a separate quotation upon request.
Request a Quote / Contact Us for TrinityX Cluster Manager
We can customize the solution to meet your specific requirements.
Please feel free to consult with us regarding system configurations or any other needs.
*Required fields
PARTNERS
