NVIDIA / deepops

Tools for building GPU clusters
BSD 3-Clause "New" or "Revised" License
1.25k stars 326 forks source link

Deploy Kubernetes and SLRUM Offline #1227

Closed jittu11 closed 2 years ago

jittu11 commented 2 years ago

Hi Team,

I need to know the feasibility of implementing DGX A 100 with deepops as a clustermanager in an air gapped environment. can we able to implement kubernetes and Slrum offline. Can i get some idea on this?

ajdecon commented 2 years ago

DeepOps has very limited support for air-gapped deployments, and we don't currently have tested procedures for either Kubernetes or Slurm in this model. We have some additional documentation on this here.

NVIDIA Bright Cluster Manager does provide air gap support, including a documented procedure for Kubernetes deployments. If you're looking for a well-tested process, I'd probably suggest using Bright for this.

jittu11 commented 2 years ago

Hi Team,

I am planning to use RHEL , can we able to implement hybrid cluster with Deepops on RHEL.