Prepare to deploy Foundry Local on Azure Local in disconnected environments

  • Applies to: Foundry Local on Azure Local

This article outlines the prerequisites and steps to download and import the Foundry Local extension expansion pack in preparation for deployment in a disconnected environment.

Important

  • Foundry Local is available in preview. Preview releases provide early access to features that are in active deployment.
  • Features, approaches, and processes can change or have limited capabilities before general availability (GA).

Prerequisites

Before you begin, complete the following checks:

  • Azure Local Disconnected Operations is installed on-premises. The minimum supported version is 2604.3.0.

  • An AKS Arc cluster on Azure Local, registered with Azure Arc as a connectedClusters resource.

    Requirement Minimum Recommended
    Worker node VM size Standard_D4s_v3 (4 vCPU / 16 GiB) Standard_D8s_v3 (8 vCPU / 32 GiB)
    Allocatable memory/node >= 14 GiB >= 28 GiB
    Worker node count 1 2+ (HA or GPU pool separation)
    Logical network Reachable from edgeartifacts ACR -

    If you plan to run multireplica vLLM deployments, reserve extra capacity for one EPP pod per ModelDeployment (about 512 MiB request and 2 GiB limit). For default behavior and tunables, see ModelDeployment and operator configuration reference.

    Don't use the az aksarc create default worker size Standard_A4_v2 (8 GiB). Use at least Standard_D4s_v3.

  • (Optional) GPU node pool (required only for GPU model variants such as *-cuda-gpu and vLLM):

    • NVIDIA DDA-passthrough SKU: Standard_NC*_A2, Standard_NC*_L4_*, Standard_NC*_L40_*, Standard_NC*_L40S_*, Standard_NC*_RTX6000Pro_*, or Tesla T4 Standard_NK*. AMD GPUs aren't supported.
    • NVIDIA mitigation INF is installed on the physical Azure Local node.
    • nvidia/k8s-device-plugin:v0.11.0 is mirrored into the edgeartifacts container registry at the path expected by the auto-deployed DaemonSet.

For disconnected architecture and bundled dependency details, see Foundry Local on Azure Local in disconnected environments overview.

Download and import the Foundry Local expansion pack

Download the Foundry Local extension expansion pack in a connected environment, transfer it to your disconnected environment, and then install it on the Azure Local Disconnected Operations machine.

  1. Download the latest Foundry Local extension expansion pack. For additional versions and release notes, see the Foundry Local Disconnected Catalog.

  2. Transfer the expansion pack to the disconnected Azure Local environment.

  3. Run the following commands on the Azure Local Disconnected Operations machine to install the expansion pack.

    Replace <PATH_TO_EXPANSION_PACK> with the local path to the expansion pack zip file and <PATH_TO_ALDO_MODULES> before you run the command.

    Import-Module "<PATH_TO_ALDO_MODULES>\Azure.Local.ExpansionPack.psm1"

$expansionPackId = Start-AldoExpansionPackUpload `
    -ExpansionPackPath "<PATH_TO_EXPANSION_PACK>"

$result = Start-AldoExpansionPackInstallation `
    -ExpansionPackId $expansionPackId `
    -Wait

When installation finishes:

  • Container images are imported into the edgeartifacts registry.
  • Model artifacts are published to the registry.
  • The Foundry Local Azure Arc extension becomes available for installation.

Verify installation

Run the following command on the Azure Local Disconnected Operations machine and confirm your expansion pack is in Installed state.

Get-ApplianceExpansionPackDetails

Next step

Deploy Foundry Local as an Azure Arc extension in a disconnected environment

Related content

  • Last updated on