How-to Guides

Task-oriented procedures for deploying, configuring, and managing Omnia clusters. Each guide follows a consistent structure: Overview, Prerequisites, Procedure, Verification, Next Steps, and Troubleshooting.

Tip

If you are new to Omnia, start with the Index tutorials first. How-to guides assume you understand Omnia's architecture and have a working OIM.

Initial Setup & Provisioning

Get Omnia installed on your OIM and provision bare-metal servers into a working cluster.

Slurm Job Scheduling

Deploy and manage Slurm-based HPC clusters, including GPU-accelerated workloads and dynamic node management.

Kubernetes Services

Deploy and configure the Kubernetes service cluster used for platform services, monitoring, and storage.

Storage

Configure shared storage for your cluster, including NFS and PowerVault block storage.

Networking

Set up high-performance interconnects for your compute fabric, including InfiniBand and RoCE.

Authentication

Configure centralized user authentication across your cluster using LDAP.

Telemetry & Monitoring

Deploy and configure the telemetry pipeline for cluster-wide metrics collection, aggregation, and visualization.

Advanced Configurations

Deploy additional packages and container images, configure Apptainer for pulling images from Pulp registries, and deploy additional repositories.

BuildStreaM (CI/CD)

Automate cluster deployment using GitLab CI/CD pipelines and the BuildStreaM catalog-driven workflow.