Air-gapped BCM + Kubernetes + Slurm installation workflow

Infrastructure & Networking Base Command Manager
1

Hi,
we are trying to deploy Bright Cluster Manager, Kubernetes, and Slurm in a fully air-gapped environment. During the process we encountered several issues related to:

  • loading required BCM packages into our local APT repository,

  • missing Bright packages (e.g., cm-docker, cm-containerd, cm-nvidia-container-toolkit) when running cm-docker-setup,

  • containerd CRI not initializing correctly without the Bright-provided packages,

  • Kubernetes Setup failing during image pull and CRI validation even with air-gap registry + ChartMuseum configured.

We want to confirm the correct/official sequence for deploying Bright in an air-gapped installation that includes:

  1. Preparing local APT repository for all required BCM packages

  2. Loading Docker/Containerd configs for air-gap

  3. Loading Kubernetes images and Helm charts

  4. Running Slurm and the remaining workload managers

Is there an officially documented step-by-step workflow for BCM + Kubernetes + Slurm in air-gapped mode?
If yes, could you guide us to the correct documentation or recommended procedure?

Thanks in advance.