kubernetes/kops · error

expected a aws.Cloud provider

Error message

expected a aws.Cloud provider

What it means

Returned immediately after the launch template specification is resolved: LaunchTemplateId was present as a field but empty (aws.ToString yielded ""). kOps needs a concrete ID to describe/compare the template during rolling updates.

Source

Thrown at pkg/resources/aws/aws.go:1438

					for _, eip := range response.Addresses {
						eipTracker := buildElasticIPResource(eip, !ownedNatGatewayIds.Has(natGatewayId), clusterName)
						resourceTrackers = append(resourceTrackers, eipTracker)
					}
				}
			}
		}
	}

	return resourceTrackers, nil
}

// DeleteAutoScalingGroupLaunchTemplate deletes
func DeleteAutoScalingGroupLaunchTemplate(cloud fi.Cloud, r *resources.Resource) error {
	ctx := context.TODO()
	c, ok := cloud.(awsup.AWSCloud)
	if !ok {
		return errors.New("expected a aws.Cloud provider")
	}
	klog.V(2).Infof("Deleting EC2 LaunchTemplate %q", r.ID)

	if _, err := c.EC2().DeleteLaunchTemplate(ctx, &ec2.DeleteLaunchTemplateInput{
		LaunchTemplateId: new(r.ID),
	}); err != nil {
		return fmt.Errorf("error deleting ec2 LaunchTemplate %q: %v", r.ID, err)
	}

	return nil
}

func DeleteELB(cloud fi.Cloud, r *resources.Resource) error {
	ctx := context.TODO()
	c := cloud.(awsup.AWSCloud)

	id := r.ID

View on GitHub (pinned to 4c8573c808)

Solutions

1. Inspect the ASG's MixedInstancesPolicy with `aws autoscaling describe-auto-scaling-groups` and fix the launch template reference
2. Re-create the ASG through `kops update cluster` so a valid launch template is attached
3. Re-run the operation if it was a transient mid-update race

Example fix

// before
id := aws.ToString(launchTemplate.LaunchTemplateId)
if id == "" {
	return "", fmt.Errorf("error finding launch template ID for autoscaling group: %s", aws.ToString(g.AutoScalingGroupName))
}
// after
if launchTemplate.LaunchTemplateId == nil || aws.ToString(launchTemplate.LaunchTemplateId) == "" {
	return "", fmt.Errorf("ASG %s has a launch template spec with empty LaunchTemplateId", aws.ToString(g.AutoScalingGroupName))
}
id := aws.ToString(launchTemplate.LaunchTemplateId)

if g.LaunchTemplate != nil && aws.ToString(g.LaunchTemplate.LaunchTemplateId) == "" {
  return fmt.Errorf("ASG %q has empty LaunchTemplateId", aws.ToString(g.AutoScalingGroupName))
}

func hasLaunchTemplateID(g *autoscalingtypes.AutoScalingGroup) bool {
  lt := g.LaunchTemplate
  if lt == nil && g.MixedInstancesPolicy != nil && g.MixedInstancesPolicy.LaunchTemplate != nil { lt = g.MixedInstancesPolicy.LaunchTemplate.LaunchTemplateSpecification }
  return lt != nil && aws.ToString(lt.LaunchTemplateId) != ""
}

err := op(ctx)
if strings.Contains(err.Error(), "error finding launch template ID") {
  // re-describe the ASG; a mid-update race often resolves on a fresh read
  time.Sleep(backoff); retry()
}

When it happens

Trigger: The ASG's LaunchTemplateSpecification exists but has an empty/nil LaunchTemplateId — typically a malformed or partially populated MixedInstancesPolicy returned by the API, or an ASG created/updated concurrently.

Common situations: Externally managed ASGs with broken mixed-instances policy; race during ASG update where the template spec is mid-mutation; custom tooling writing incomplete ASG definitions.

Related errors

• DIGITALOCEAN_ACCESS_TOKEN is required
• errShutdown
• could not find one of launch template or mixed instances pol
• DIGITALOCEAN_ACCESS_TOKEN is required
• the image for the hook exec action not set

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/659f3e364f9e93a6. Report an issue: GitHub.