derailed/k9s · error

unable to locate node %s

Error message

unable to locate node %s

What it means

Node.Get lists all nodes and scans for one whose name matches path; when no entry matches it returns this error. The List call succeeded, so this is strictly a name-mismatch / not-found-after-list failure, not an RBAC or connectivity problem.

Source

Thrown at internal/dao/node.go:147

	return nil
}

// Get returns a node resource.
func (n *Node) Get(ctx context.Context, path string) (runtime.Object, error) {
	oo, err := n.Resource.List(ctx, "")
	if err != nil {
		return nil, err
	}

	var raw *unstructured.Unstructured
	for _, o := range oo {
		if u, ok := o.(*unstructured.Unstructured); ok && u.GetName() == path {
			raw = u
		}
	}
	if raw == nil {
		return nil, fmt.Errorf("unable to locate node %s", path)
	}

	var nmx *mv1beta1.NodeMetrics
	if withMx, ok := ctx.Value(internal.KeyWithMetrics).(bool); ok && withMx {
		nmx, _ = client.DialMetrics(n.Client()).FetchNodeMetrics(ctx, path)
	}

	podCount := -1
	if shouldCountPods, _ := ctx.Value(internal.KeyPodCounting).(bool); shouldCountPods {
		if pp, err := n.GetPods(path); err == nil {
			podCount = len(pp)
		}
	}

	return &render.NodeWithMetrics{Raw: raw, MX: nmx, PodCount: podCount}, nil
}

// List returns a collection of node resources.

View on GitHub (pinned to 2d3ccc6ba2)

Solutions

1. Verify the node exists in the same context: kubectl get nodes | grep <name>
2. Refresh/relist (reopen the view or restart the session) and retry once
3. Check the kubeconfig context — the name may belong to another cluster

nodes, err := dao.FetchNodes(ctx, factory, "")
if err != nil { return err }
found := false
for _, nn := range nodes.Items {
    if nn.Name == name { found = true; break }
}
if !found { return fmt.Errorf("node %q not present in cluster", name) }

if _, err := nodeDAO.Get(ctx, path); err != nil {
    if strings.Contains(err.Error(), "unable to locate node") {
        // one refresh + retry; after that surface 'node not found' to the user
    }
}

When it happens

Trigger: Node deleted (or recreated under a new name) between the list and the lookup; stale or mistyped node name; the DAO cache not yet synced after cluster changes.

Common situations: Autocomplete or saved reference from a previous session/cluster; node removed while a k9s session stayed open; kubeconfig context pointing at a different cluster that does not have that node.

Related errors

• no connection to cached dial
• node is cordoned
• failed to locate pod %q: %w
• node is already cordoned
• node is already uncordoned

AI-assisted analysis of derailed/k9s@2d3ccc6ba2 (2026-08-15). Data as JSON: /api/errors/13acd7fda2fa5c1e. Report an issue: GitHub.