Install Omnia on Ubuntu clusters

Note

• Slurm installation is not supported on Ubuntu clusters.

• FreeIPA configuration is not supported on Ubuntu clusters.

• Prerequisites
• Space requirements for the OIM running on Ubuntu OS
• Step 1: Execute prereq.sh
• Step 2: Create Local repositories for the cluster
  • Prerequisites
  • Hostname requirements
  • Input parameters for Local Repositories
  • Configure specific local repositories
  • Execute local repo playbook
• Step 3: Discover and provision the cluster
  • Before you run the provision tool
  • Discovery Mechanisms
  • Input parameters for the provision tool
  • Provisioning the cluster
  • Checking node status
• Step 4: View node inventory
• Step 5: Configure the cluster
  • Before you build clusters
  • Input parameters for the cluster
  • Building an Omnia Cluster
• Step 6: Install AI tools
  • Setup Jupyterhub
  • Setup Kubeflow
  • Deploying Dell Enterprise Pretrained model on the cluster
  • Setup DeepSpeed for clusters containing Intel Gaudi accelerators
  • Setup vLLM
  • Setup PyTorch
  • Setup TensorFlow
  • Setup Kserve
• Advanced configurations for Ubuntu clusters
  • Configure additional NICs and specify Kernel Parameters on the nodes
  • Configuring custom repositories
  • Configuring UCX and OpenMPI on the cluster
  • Kubernetes plugin for RoCE NIC
  • Granting Kubernetes access
  • Configuring switches
  • Configuring Storage
  • Alternate method to install the AMD ROCm platform
  • Alternate method to install the Intel Gaudi Software Stack and Driver
  • Deploy CSI drivers for Dell PowerScale storage solutions

If you have any feedback about Omnia documentation, please reach out at omnia.readme@dell.com.