kubernetes/kops · error

failed to find network for loadbalancer %q

Error message

failed to find network for loadbalancer %q

What it means

In RenderHetzner, when creating a NEW load balancer (a == nil), the task requires e.Network (the Hetzner network the LB should join) because the create call passes a network ID and targets use UsePrivateIP. This error is thrown when the LoadBalancer task has no Network set, so kOps cannot resolve the networkID for hcloud LoadBalancerCreateOpts.

Source

Thrown at upup/pkg/fi/cloudup/hetznertasks/loadbalancer.go:235

		}
		if len(e.Services) == 0 {
			return fi.RequiredField("Services")
		}
		if e.Target == "" {
			return fi.RequiredField("Target")
		}
	}
	return nil
}

func (_ *LoadBalancer) RenderHetzner(t *hetzner.HetznerAPITarget, a, e, changes *LoadBalancer) error {
	ctx := context.TODO()
	actionClient := t.Cloud.ActionClient()
	client := t.Cloud.LoadBalancerClient()

	if a == nil {
		if e.Network == nil {
			return fmt.Errorf("failed to find network for loadbalancer %q", fi.ValueOf(e.Name))
		}

		networkID, err := strconv.ParseInt(fi.ValueOf(e.Network.ID), 10, 64)
		if err != nil {
			return fmt.Errorf("failed to convert network ID %q to int: %w", fi.ValueOf(e.Network.ID), err)
		}

		opts := hcloud.LoadBalancerCreateOpts{
			Name: fi.ValueOf(e.Name),
			LoadBalancerType: &hcloud.LoadBalancerType{
				Name: e.Type,
			},
			Algorithm: &hcloud.LoadBalancerAlgorithm{
				Type: hcloud.LoadBalancerAlgorithmTypeRoundRobin,
			},
			Location: &hcloud.Location{
				Name: e.Location,
			},

View on GitHub (pinned to 4c8573c808)

Solutions

1. Add the Hetzner network configuration to the cluster spec so the LoadBalancer task gets a Network, then re-run kops update cluster --yes.
2. Ensure the cluster's network task is created before the load balancer (check the task builder wiring so e.Network is populated via dependencies).
3. If a networkless LB is truly desired, the create path must be changed to omit the Network/UsePrivateIP fields — this is not supported by the current kOps Hetzner implementation, so do not bypass it.

Example fix

// before: cluster.yaml without hetzner network
spec:
  networkID: ""   # no network -> LoadBalancer task has Network == nil
// after: define the network in the cluster spec so the LB task resolves it
spec:
  network:
    networkID: "1234567"
    zone: fsn1-dc14

if e.Network == nil || fi.ValueOf(e.Network.ID) == "" {
    return fmt.Errorf("hetzner load balancer %q requires a network; set networkID in the cluster spec", fi.ValueOf(e.Name))
}
if _, err := strconv.ParseInt(fi.ValueOf(e.Network.ID), 10, 64); err != nil {
    return fmt.Errorf("network ID %q is not numeric", fi.ValueOf(e.Network.ID))
}

func hasNetwork(e *LoadBalancer) bool {
    return e.Network != nil && fi.ValueOf(e.Network.ID) != ""
}

When it happens

Trigger: A Hetzner cluster is applied without a network defined in the cluster spec, so the LoadBalancer task's Network field is nil at render time; or the task graph omitted the Network dependency when constructing the LoadBalancer task.

Common situations: Cluster spec missing spec.networking.hetzner / network configuration; hand-edited or partial cluster manifest; version mismatch where the load balancer was introduced requiring a network the cluster never had.

Understand the failure class

Background: "X is required", "must be set", "cannot be empty": the missing-required-config error family, from Vertex AI project/location to WeChat keys — this error's family across 18 libraries.

Related errors

• failed to find load-balancer %q public address
• failed to find load-balancer %q private address
• %s is required
• expected exactly one subnet with GCE IP Aliases
• failed to get load balancers matching label selector %q: %w

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/12c27939be181525. Report an issue: GitHub.