kubernetes/kops · error

error creating loadbalancer: %v

Error message

error creating loadbalancer: %v

What it means

Within createLB, loadbalancers.Create is retried via vfs.RetryWithBackoff; every failure from the Octavia API is wrapped as 'error creating loadbalancer: %v' and retried until writeBackoff gives up. The message carries the underlying API error (validation, quota, conflict, outage).

Source

Thrown at upup/pkg/fi/cloudup/openstack/loadbalancer.go:221

	} else {
		return wait.ErrWaitTimeout
	}
}

func (c *openstackCloud) CreateLB(opt loadbalancers.CreateOptsBuilder) (*loadbalancers.LoadBalancer, error) {
	return createLB(c, opt)
}

func createLB(c OpenstackCloud, opt loadbalancers.CreateOptsBuilder) (*loadbalancers.LoadBalancer, error) {
	if c.LoadBalancerClient() == nil {
		return nil, fmt.Errorf("loadbalancer support not available in this deployment")
	}

	var i *loadbalancers.LoadBalancer
	done, err := vfs.RetryWithBackoff(writeBackoff, func() (bool, error) {
		v, err := loadbalancers.Create(context.TODO(), c.LoadBalancerClient(), opt).Extract()
		if err != nil {
			return false, fmt.Errorf("error creating loadbalancer: %v", err)
		}
		i = v
		return true, nil
	})
	if err != nil {
		return i, err
	} else if done {
		return i, nil
	} else {
		return i, wait.ErrWaitTimeout
	}
}

func (c *openstackCloud) GetLB(loadbalancerID string) (lb *loadbalancers.LoadBalancer, err error) {
	return getLB(c, loadbalancerID)
}

func getLB(c OpenstackCloud, loadbalancerID string) (lb *loadbalancers.LoadBalancer, err error) {

View on GitHub (pinned to 4c8573c808)

Solutions

1. Read the wrapped error: 403/quota — raise the project's loadbalancer quota; 404 — fix vip subnet/network IDs in the spec
2. Validate CreateOpts against the deployed Octavia version (flavor, provider, vip_subnet_id) and correct the cluster spec
3. Check Octavia service health (amphora capacity, API logs) if errors are 5xx and retry
4. Confirm the configured project/network can host the VIP and has ports available
5. Retry the operation after transient failures — backoff is already applied, so persistence implies a config/quota issue

// pre-validate CreateOpts against the environment
if opt.GetVipSubnetID() == "" { return fmt.Errorf("vip_subnet_id is required") }
if flavorID != "" {
	if _, err := flavors.Get(context.TODO(), client, flavorID).Extract(); err != nil {
		return fmt.Errorf("flavor %s does not exist in this Octavia", flavorID)
	}
}
// also check quota: os_quota.versions loadbalancer limit

done, err := vfs.RetryWithBackoff(writeBackoff, func() (bool, error) {
	v, err := loadbalancers.Create(context.TODO(), client, opt).Extract()
	if err != nil {
		var gerr gophercloud.ErrUnexpectedResponseCode
		if errors.As(err, &gerr) && (gerr.Actual == 403 || gerr.Actual == 404) {
			return false, err // config/quota problem: do not blind-retry
		}
		return false, fmt.Errorf("error creating loadbalancer: %v", err) // transient: retry
	}
	return true, nil
})

When it happens

Trigger: Octavia create rejected: invalid CreateOpts (bad vip_subnet_id/flavor), quota exceeded for loadbalancers, subnet/vip not found, 409 on provisioning status, or API 5xx during service provisioning for a LoadBalancer-type Kubernetes service.

Common situations: Kubernetes service of type LoadBalancer stuck Pending because Octavia quota (loadbalancer count) is exhausted; wrong subnet/network in the cluster spec; Octavia flavor_id referencing a nonexistent flavor; API outage during scale-up.

Related errors

• error deleting loadbalancer: %v
• failed to list loadbalancers: %s
• failed to list members: %v
• failed to list pools: %v
• failed to list listeners: %v

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/d29ad327deedf3c4. Report an issue: GitHub.