kubernetes/kops · error

error listing Instances: %v

Error message

error listing Instances: %v

What it means

In the GCE Instance task's Find, kOps fetches the instance with Instances().Get. If the call fails with anything other than a NotFound, the error is wrapped as 'error listing Instances'. It means the instance lookup itself failed at the GCE API level, not that the instance is absent.

Source

Thrown at upup/pkg/fi/cloudup/gcetasks/instance.go:74

	metadataFingerprint string
}

var _ fi.CompareWithID = (*Instance)(nil)

func (e *Instance) CompareWithID() *string {
	return e.Name
}

func (e *Instance) Find(c *fi.CloudupContext) (*Instance, error) {
	cloud := c.T.Cloud.(gce.GCECloud)

	r, err := cloud.Compute().Instances().Get(cloud.Project(), *e.Zone, *e.Name)
	if err != nil {
		if gce.IsNotFound(err) {
			return nil, nil
		}
		return nil, fmt.Errorf("error listing Instances: %v", err)
	}

	actual := &Instance{}
	actual.Name = &r.Name
	actual.Tags = append(actual.Tags, r.Tags.Items...)
	actual.Zone = new(lastComponent(r.Zone))
	actual.MachineType = new(lastComponent(r.MachineType))
	actual.CanIPForward = &r.CanIpForward
	if r.Scheduling != nil {
		actual.Preemptible = &r.Scheduling.Preemptible
	}
	if len(r.NetworkInterfaces) != 0 {
		ni := r.NetworkInterfaces[0]
		actual.Network = &Network{Name: new(lastComponent(ni.Network))}
		actual.StackType = &ni.StackType
		if len(ni.AccessConfigs) != 0 {
			ac := ni.AccessConfigs[0]
			if ac.NatIP != "" {

View on GitHub (pinned to 4c8573c808)

Solutions

1. Read the wrapped %v error to identify the underlying GCE API failure
2. Verify the zone in the instance spec/state exists: `gcloud compute zones list`
3. Enable the Compute Engine API and check IAM: service account needs compute.viewer (compute.instances.get)
4. Retry if the error is transient (429/5xx)

// Verify the zone and API access before the Get call
cmd := exec.Command("gcloud", "compute", "instances", "describe", *e.Name,
	"--zone", *e.Zone, "--project", cloud.Project(), "--format=value(name)")
if err := cmd.Run(); err != nil {
	return fmt.Errorf("precheck: instance %s not accessible in zone %s: %w", *e.Name, *e.Zone, err)
}

func isNotFound(err error) bool { return err != nil && gce.IsNotFound(err) }
func isPermissionDenied(err error) bool {
	var ge *googleapi.Error
	return errors.As(err, &ge) && ge.Code == 403
}

r, err := cloud.Compute().Instances().Get(cloud.Project(), *e.Zone, *e.Name)
if err != nil {
	if gce.IsNotFound(err) { return nil, nil }
	var ge *googleapi.Error
	if errors.As(err, &ge) && (ge.Code == 429 || ge.Code >= 500) {
		return nil, retry.Wrapf(err, "transient; retry with backoff")
	}
	return nil, fmt.Errorf("error listing Instances: %v", err)
}

When it happens

Trigger: Instances().Get(project, zone, name) returns a non-NotFound error: invalid/unknown zone, permission denied on the compute API, API disabled, network failure, or quota/billing issues.

Common situations: Zone removed or mis-set in the state vs. current project, Compute Engine API disabled for the project, service account lacking compute.instances.get IAM permission, or transient API outages during `kops update cluster --refresh`.

Related errors

• error fetching GCE instance: %w
• error fetching GCE instance group template %q: %v
• error fetching GCE managed instance group %q: %v
• error getting ForwardingRule %q: %w
• error getting TargetPool %q: %w

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/2a01f13b2a097b43. Report an issue: GitHub.