kubernetes/kops · error

error storing InstanceGroup: %v

Error message

error storing InstanceGroup: %v

What it means

Once validated, the InstanceGroup is persisted via clientset.InstanceGroupsFor(cluster).Create. If the kOps API (state store backend, e.g. S3-backed) fails to create the object, the underlying error is wrapped with this message. At this point the group definition is valid but was not saved, so the cluster state is unchanged.

Source

Thrown at cmd/kops/create_instancegroup.go:295

		if err != nil {
			return fmt.Errorf("error parsing yaml: %v", err)
		}
		group, ok := obj.(*kopsapi.InstanceGroup)
		if !ok {
			return fmt.Errorf("unexpected object type: %T", obj)
		}

		err = validation.CrossValidateInstanceGroup(group, cluster, cloud, true).ToAggregate()
		if err != nil {
			return err
		}

		ig = group
	}

	_, err = clientset.InstanceGroupsFor(cluster).Create(ctx, ig, metav1.CreateOptions{})
	if err != nil {
		return fmt.Errorf("error storing InstanceGroup: %v", err)
	}

	return nil
}

func completeClusterSubnet(f commandutils.Factory, excludeSubnets *[]string) func(cmd *cobra.Command, args []string, toComplete string) ([]string, cobra.ShellCompDirective) {
	return func(cmd *cobra.Command, args []string, toComplete string) ([]string, cobra.ShellCompDirective) {
		ctx := cmd.Context()

		commandutils.ConfigureKlogForCompletion()

		cluster, _, completions, directive := GetClusterForCompletion(ctx, f, nil)
		if cluster == nil {
			return completions, directive
		}

		if len(args) > 1 {
			return commandutils.CompletionError("too many arguments", nil)

View on GitHub (pinned to 4c8573c808)

Solutions

1. Read the wrapped error: if it says already exists, the InstanceGroup name is taken — use `kops edit instancegroup` or delete and recreate
2. Verify KOPS_STATE_STORE is correct and the backend (S3/GCS/etc.) is reachable with valid credentials
3. Check IAM/bucket permissions for write access to the state store path
4. Retry after restoring network connectivity to the cloud storage backend
5. Run `kops get instancegroups` to confirm current state before retrying

Example fix

// before
kops create instancegroup nodes --role Node  # AlreadyExists
// after
kops edit instancegroup nodes  # or kops delete instancegroup nodes first

kops get cluster "$CLUSTER" >/dev/null && kops get instancegroup -n "$CLUSTER" "$IG_NAME" 2>/dev/null && echo exists
# also check state store reachability/credentials beforehand

if err != nil && strings.Contains(err.Error(), "error storing InstanceGroup") {
	if strings.Contains(err.Error(), "already exists") { /* switch to edit/update path */ } else { /* retry with backoff after checking state store access */ }
}

When it happens

Trigger: Calling Create on the InstanceGroup clientset when the state store is unreachable or read-only, the object already exists (AlreadyExists conflict), permissions are missing, or the context is cancelled.

Common situations: S3 bucket permissions changed or AWS credentials expired; kops state store (KOPS_STATE_STORE) pointing at a wrong/missing bucket; the InstanceGroup name already exists (use `kops update`/`kops edit` instead); network outage to the backend.

Related errors

• adding encryptionconfig secret: %v
• updating encryptionconfig secret: %v
• error creating cluster: %v
• error reading state store: %v
• error reading %s %q: %v

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/a7c1c5733a526e83. Report an issue: GitHub.