kubernetes/kops · error

error during Disk creation: %v

Error message

error during Disk creation: %v

What it means

kOps' GCE Disk task throws this when the long-running Google Compute API Insert operation for a persistent disk completes but WaitForOp reports the operation itself failed. The disk creation API call was accepted, but GCP returned an operation error during asynchronous provisioning.

Source

Thrown at upup/pkg/fi/cloudup/gcetasks/disk.go:139

		SizeGb: *e.SizeGB,
		Type:   typeURL,
	}

	if e.VolumeIops != nil {
		disk.ProvisionedIops = *e.VolumeIops
	}
	if e.VolumeThroughput != nil {
		disk.ProvisionedThroughput = *e.VolumeThroughput
	}

	if a == nil {
		op, err := cloud.Compute().Disks().Insert(t.Cloud.Project(), *e.Zone, disk)
		if err != nil {
			return fmt.Errorf("error creating Disk: %v", err)
		}
		err = cloud.WaitForOp(op)
		if err != nil {
			return fmt.Errorf("error during Disk creation: %v", err)
		}
	}

	if changes.Labels != nil {
		d, err := cloud.Compute().Disks().Get(t.Cloud.Project(), *e.Zone, disk.Name)
		if err != nil {
			return fmt.Errorf("error reading created Disk: %v", err)
		}

		labelsRequest := &compute.ZoneSetLabelsRequest{
			LabelFingerprint: d.LabelFingerprint,
			Labels:           make(map[string]string),
		}
		// Danger: labels replace tags on instances; but thankfully volumes don't have tags
		//for _, k := range d.Tags {
		//	labelsRequest.Labels[k] = ""
		//}
		maps.Copy(labelsRequest.Labels, d.Labels)

View on GitHub (pinned to 4c8573c808)

Solutions

1. Inspect the wrapped %v error for the GCP operation error detail (often quota or zone capacity)
2. Check persistent disk quota in the target zone via gcloud compute project-info describe / console
3. Verify disk size and type are supported in the zone
4. Retry the kops update once transient GCP issues resolve
5. Choose a different zone if capacity/quota is persistently exhausted

Example fix

// before
op, err := cloud.Compute().Disks().Insert(t.Cloud.Project(), *e.Zone, disk)
// after: check quota/zone from the wrapped op error, e.g.
// gcloud compute disks create <name> --size=<size> --zone=<zone> --type=<type>
// to reproduce the raw GCP error before retrying

// pre-check quota and zone validity
desc, _ := computeService.Zones.Get(project, zone).Do()
if desc == nil || desc.Status != "UP" { return fmt.Errorf("zone %s unavailable", zone) }

if err := fi.WaitForOp(...); err != nil {
  if gcp.IsQuotaExceeded(err) { /* backoff and retry, or request quota increase */ }
  return fmt.Errorf("disk create op failed: %w", err)
}

When it happens

Trigger: RenderGCE creates a new disk via cloud.Compute().Disks().Insert(project, zone, disk); the returned op is passed to cloud.WaitForOp(op), and the operation ends with an error state (e.g. quota exceeded, zone unavailable, invalid disk spec).

Common situations: GCP regional persistent disk quota exhausted in the zone; disk size/ type invalid for the zone; zone capacity issues; project-level disk limits; transient GCP API failures during cluster bring-up.

Related errors

• error reading created Disk: %v
• error setting labels on created Disk: %v
• cannot apply changes to Disk: %v
• creating gce IPAM controller: %w
• error getting ingress status: %v

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/1ee1022d3526c651. Report an issue: GitHub.