kubernetes/kops · error

unable to determine machine architecture for InstanceGroup %

Error message

unable to determine machine architecture for InstanceGroup %q: %v

What it means

When ig.Spec.Image is empty, kOps must pick a default image matching the machine's CPU architecture. It calls MachineArchitecture(cloud, ig.Spec.MachineType); if that lookup fails (unsupported machine type string or cloud API error), this error is returned.

Source

Thrown at upup/pkg/fi/cloudup/populate_instancegroup_spec.go:152

			ig.Spec.MachineType, err = defaultMachineType(cloud, cluster, ig)
			if err != nil {
				return nil, fmt.Errorf("error assigning default machine type for nodes: %v", err)
			}
		}
		if ig.Spec.Manager != kops.InstanceManagerKarpenter {
			if ig.Spec.MinSize == nil {
				ig.Spec.MinSize = new(int32(2))
			}
			if ig.Spec.MaxSize == nil {
				ig.Spec.MaxSize = new(int32(2))
			}
		}
	}

	if ig.Spec.Image == "" {
		architecture, err := MachineArchitecture(cloud, ig.Spec.MachineType)
		if err != nil {
			return nil, fmt.Errorf("unable to determine machine architecture for InstanceGroup %q: %v", ig.ObjectMeta.Name, err)
		}
		ig.Spec.Image, err = defaultImage(cluster, channel, architecture)
		if err != nil {
			return nil, fmt.Errorf("unable to determine default image for instance group %q: %v", ig.ObjectMeta.Name, err)
		}
	}

	if ig.Spec.Tenancy != "" && ig.Spec.Tenancy != "default" {
		switch cluster.GetCloudProvider() {
		case kops.CloudProviderAWS:
			if _, ok := awsDedicatedInstanceExceptions[ig.Spec.MachineType]; ok {
				return nil, fmt.Errorf("invalid dedicated instance type: %s", ig.Spec.MachineType)
			}
		default:
			klog.Warning("Trying to set tenancy on non-AWS environment")
		}
	}

View on GitHub (pinned to 4c8573c808)

Solutions

1. Set spec.image explicitly so architecture inference is skipped
2. Correct the spec.machineType value (verify it exists for your cloud/region)
3. Upgrade kOps to gain support for newer instance families/architectures
4. Inspect the wrapped error for the underlying cloud API failure (credentials, region)

Example fix

// before
spec:
  machineType: t3.medium
// after
spec:
  machineType: t3.medium
  image: ubuntu/images/hvm-ssd/ubuntu-jammy-22.04-amd64-server-*

// Validate machine type + image pair before populating
if ig.Spec.Image == "" {
    // only omit image when machineType is a known, supported family
    known := []string{"t3.", "m5.", "c5.", "n1-", "Standard_D"}
    ok := false
    for _, p := range known {
        if strings.HasPrefix(ig.Spec.MachineType, p) { ok = true; break }
    }
    if !ok { return fmt.Errorf("set spec.image explicitly for machineType %q", ig.Spec.MachineType) }
}

When it happens

Trigger: Creating/updating an InstanceGroup with no spec.image set and a machine type whose architecture cannot be resolved — e.g. unknown/typo'd instance type, or cloud (GCE/Azure/etc.) whose MachineArchitecture implementation cannot parse or query the type.

Common situations: Typo in machineType; machine type not valid for the cloud provider; custom machine types the architecture mapper doesn't recognize; older kOps versions lacking support for new instance families (e.g. ARM Graviton types).

Related errors

• did not find owner for node %q
• invalid InstanceGroup name: %v
• error building InstanceGroup from CAPI Machine: %w
• must specify %q label with cluster name to create instanceGr
• cluster %q not found

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/0898b5b59c61cf94. Report an issue: GitHub.