abiosoft/colima · warning

error retrieving containerd volumes: %w

Error message

error retrieving containerd volumes: %w

What it means

After listing containerd namespaces, the poll runs fetchVolumes("sudo", "nerdctl", "--namespace", ns) for each namespace (nerdctl ps -q then nerdctl inspect). Any per-namespace failure wraps as 'error retrieving containerd volumes' and is logged/retried by the poll goroutine; other namespaces are skipped for that cycle.

Source

Thrown at daemon/process/inotify/volumes.go:48

			if err != nil {
				return nil, fmt.Errorf("error fetching docker volumes: %w", err)
			}
			return vols, nil

		case containerd.Name:
			var namespaces []string
			out, err := f.guest.RunOutput("sudo", "nerdctl", "namespace", "list", "-q")
			if err != nil {
				return nil, fmt.Errorf("error retrieving containerd namespaces: %w", err)
			}
			if out != "" {
				namespaces = strings.Fields(out)
			}

			for _, ns := range namespaces {
				v, err := f.fetchVolumes("sudo", "nerdctl", "--namespace", ns)
				if err != nil {
					return nil, fmt.Errorf("error retrieving containerd volumes: %w", err)
				}
				if len(v) > 0 {
					vols = append(vols, v...)
				}
			}

			return vols, nil
		}

		return nil, nil
	}

	go func() {
		for {
			select {
			case <-ctx.Done():
				log.Trace("stop signal received")
				err := ctx.Err()

View on GitHub (pinned to c3a5f9184d)

Solutions

1. wait for the next poll cycle — transient races with exiting containers heal
2. identify the failing namespace: `colima ssh -- sudo nerdctl --namespace <ns> ps -a`
3. restart colima if the guest runtime is wedged and every cycle fails

if vols, err := fetch(); err != nil {
    if strings.Contains(err.Error(), "error retrieving containerd volumes") {
        // per-namespace failure (often k8s.io churn) — next tick retries
    }
}

When it happens

Trigger: one namespace in a broken state (e.g. leftover k8s.io state); nerdctl failing only for a specific namespace; container ids vanishing between ps and inspect within that namespace.

Common situations: kubernetes-style workloads churning pods in the k8s.io namespace; partially deleted containerd state.

Related errors

• error retrieving containerd namespaces: %w
• nerdctl only supports %s runtime
• error fetching docker volumes: %w
• error starting %s: %w
• error initiating container runtime: %w

AI-assisted analysis of abiosoft/colima@c3a5f9184d (2026-08-15). Data as JSON: /api/errors/45d8cca6b98b06e6. Report an issue: GitHub.