lima-vm/lima · error

cannot safely prune while instance %#q is being used: %w

Error message

cannot safely prune while instance %#q is being used: %w

What it means

knownLocations enumerates instances to decide what prune may safely delete. If store.Inspect on a listed instance returns an error that is os.ErrNotExist, the listing is stale relative to the disk, so prune aborts with "cannot safely prune while instance <name> is being used" rather than deleting data based on inconsistent state. Other inspect errors are returned unwrapped.

Source

Thrown at cmd/limactl/prune.go:86

	}

	// Sweep leftover partial downloads (data.part / data.tmp.*) from the kept entries.
	return downloader.RemoveStaleTempFiles(opt)
}

func knownLocations(ctx context.Context) (map[string]limatype.File, error) {
	locations := make(map[string]limatype.File)

	// Collect locations from instances
	instances, err := store.Instances()
	if err != nil {
		return nil, err
	}
	for _, instanceName := range instances {
		instance, err := store.Inspect(ctx, instanceName)
		if err != nil {
			if errors.Is(err, os.ErrNotExist) {
				return nil, fmt.Errorf("cannot safely prune while instance %#q is being used: %w", instanceName, err)
			}
			return nil, err
		}
		if instance.Errors != nil {
			logrus.Warnf("skipping instance %#q because it has errors: %v", instanceName, instance.Errors)
			continue
		}
		maps.Copy(locations, locationsFromLimaYAML(instance.Config))
	}

	// Collect locations from templates
	templates, err := templatestore.Templates()
	if err != nil {
		return nil, err
	}
	for _, t := range templates {
		tmpl, err := limatmpl.Read(ctx, "", t.Location)
		if err != nil {

View on GitHub (pinned to dd909d0973)

Solutions

1. Re-run `limactl prune` once no other limactl commands are executing
2. Run `limactl list` to refresh/verify consistent instance state, then prune
3. Remove the stale/ghost instance reference (e.g. finish the interrupted delete) and retry

Example fix

// before
limactl prune &  limactl delete myvm --force   # racy
// after
limactl delete myvm --force
limactl prune

# ensure no other limactl operation is in flight
pgrep -f 'limactl (delete|prune)' && { echo 'lima operation in progress' >&2; exit 1; }

if err := pruneCmd.Run(); err != nil {
    if strings.Contains(err.Error(), "cannot safely prune while instance") {
        time.Sleep(2 * time.Second)
        // retry prune once concurrent operations have finished
    }
}

When it happens

Trigger: An instance directory listed by store.Instances() disappeared (concurrent `limactl delete`, external rm) between listing and inspection while running `limactl prune`.

Common situations: Running prune in parallel with instance deletion; manually removing ~/.lima/instance dirs while Lima commands run; race with another limactl process in CI.

Related errors

• instance %q not found
• instance %#q not found
• failed to delete instance %#q: %w
• instance %#q directory %#q exists but its %#q could not be r
• failed to inspect instance %#q: %w

AI-assisted analysis of lima-vm/lima@dd909d0973 (2026-09-01). Data as JSON: /api/errors/a6a5b6b379b43c90. Report an issue: GitHub.