kubernetes/kops · error

getting IP of server %s for load-balancer's back-end: %w

Error message

getting IP of server %s for load-balancer's back-end: %w

What it means

kOps wraps failures from `scwCloud.GetServerIP(server.ID, server.Zone)`, which resolves a control-plane server's private IP for the LB back-end, when iterating over cluster servers (workers are skipped). Thrown from getControlPlanesIPs, called by RenderScw.

Source

Thrown at upup/pkg/fi/cloudup/scalewaytasks/lb_backend.go:257

func (l *LBBackend) TerraformLink() *terraformWriter.Literal {
	return terraformWriter.LiteralProperty("scaleway_lb_backend", fi.ValueOf(l.Name), "id")
}

func getControlPlanesIPs(scwCloud scaleway.ScwCloud, lb *LoadBalancer, zone scw.Zone) ([]string, error) {
	var controlPlanePrivateIPs []string

	servers, err := scwCloud.GetClusterServers(scwCloud.ClusterName(lb.Tags), nil)
	if err != nil {
		return nil, fmt.Errorf("getting cluster servers for load-balancer's back-end: %w", err)
	}

	for _, server := range servers {
		if role := scaleway.InstanceRoleFromTags(server.Tags); role == scaleway.TagRoleWorker {
			continue
		}
		ip, err := scwCloud.GetServerIP(server.ID, server.Zone)
		if err != nil {
			return nil, fmt.Errorf("getting IP of server %s for load-balancer's back-end: %w", server.Name, err)
		}
		controlPlanePrivateIPs = append(controlPlanePrivateIPs, ip)
	}

	return controlPlanePrivateIPs, nil
}

View on GitHub (pinned to 4c8573c808)

Solutions

1. Rerun `kops update cluster` so the server list is refreshed against current instances
2. Check the named server exists in the Scaleway console and is in 'running' state with an IP assigned
3. Verify the server's zone matches the cloud zone configuration
4. Add retry/backoff if rate-limit (429) errors are reported in the wrapped error
5. If a deleted instance persistently appears, fix stale instance records or re-sync the cluster

Example fix

// before: hard fail on any server
ip, err := scwCloud.GetServerIP(server.ID, server.Zone)
if err != nil { return nil, fmt.Errorf("getting IP of server %s ...: %w", server.Name, err) }
// after: skip stale servers without an IP yet
ip, err := scwCloud.GetServerIP(server.ID, server.Zone)
if err != nil {
    klog.Warningf("skipping server %s: %v", server.Name, err)
    continue
}

// skip servers that cannot have an IP yet
if server.State != instance.ServerStateRunning { klog.Warningf("server %s not running (%s)", server.Name, server.State); continue }
if server.Zone == "" { return fmt.Errorf("server %s has no zone", server.Name) }

if len(controlPlanePrivateIPs) == 0 {
    return fmt.Errorf("no control-plane IPs resolved: last error: %w", err)
}

When it happens

Trigger: For each non-worker server from GetClusterServers, GetServerIP is called; it fails because the server was deleted between list and fetch (stale entry), the zone is wrong, or the GetServer API call errors (auth, outage, rate limit).

Common situations: Control-plane instance deleted/replaced concurrently (stale server ID); server not yet assigned an IP (still provisioning); wrong zone passed; API rate-limiting during large clusters.

Related errors

• updating back-end server IPs for load-balancer %s: %w
• getting cluster servers for load-balancer's back-end: %w
• listing IPs for deletion: %w
• failed to delete instance IP %s: %w
• error getting Instance: %v

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/81e1155a642cd6e9. Report an issue: GitHub.