lima-vm/lima · error

failed to create mount point %#q: %w

Error message

failed to create mount point %#q: %w

What it means

Before mounting the Data volume, patchWriteGuestFiles creates the instance's mount point directory (<instance dir>/mnt) with os.MkdirAll. If directory creation fails (permission problem, path is a file, read-only volume), the OS error is wrapped with this message including the path.

Source

Thrown at pkg/guestpatch/macos/macos_darwin.go:71

	attached, err := attachImageWithRetry(ctx, disk, 3)
	if err != nil {
		return fmt.Errorf("failed to attach disk: %w", err)
	}
	if attached == nil || attached.Data == "" {
		return errors.New("failed to find data slice in attached disk")
	}
	dataDevPath := "/dev/" + attached.Data
	defer func() {
		// Detaching the data slice is enough to detach the whole ASIF.
		if err := asifutil.DetachASIF(dataDevPath); err != nil {
			logrus.WithError(err).Warnf("failed to detach %#q (%#q)", dataDevPath, disk)
		}
	}()

	instDir := filepath.Dir(disk)
	mnt := filepath.Join(instDir, filenames.MntDir)
	if err := os.MkdirAll(mnt, 0o755); err != nil {
		return fmt.Errorf("failed to create mount point %#q: %w", mnt, err)
	}
	defer func() {
		if err := os.Remove(mnt); err != nil {
			logrus.WithError(err).Warnf("failed to remove mount point %#q", mnt)
		}
	}()
	return writeGuestFiles(ctx, dataDevPath, mnt)
}

// patchFixOwnership attaches the disk image and patches APFS inode
// records on the raw container device to set root:wheel ownership.
func patchFixOwnership(ctx context.Context, disk string) error {
	attached, err := attachImageWithRetry(ctx, disk, 3)
	if err != nil {
		return fmt.Errorf("failed to attach disk for ownership fix: %w", err)
	}
	if attached == nil || attached.Data == "" {
		return errors.New("failed to find data slice in attached disk")

View on GitHub (pinned to dd909d0973)

Solutions

1. Check the wrapped OS error (EACCES/ENOTDIR/EEXIST-as-file) and the quoted path.
2. Verify write permission on the instance directory and correct ownership (`ls -ld`).
3. Remove any regular file occupying the mnt path.
4. Ensure the volume holding LIMA_HOME is writable (not read-only mounted).

Example fix

// before: mnt path occupied by a file
$ ls ~/.lima/inst/mnt -> regular file
// after
$ rm ~/.lima/inst/mnt   # remove stale file
$ chmod u+w ~/.lima/inst

// pre-validate the mount point path is creatable
mnt := filepath.Join(filepath.Dir(disk), "mnt")
if fi, err := os.Lstat(mnt); err == nil && !fi.IsDir() {
	return fmt.Errorf("%s exists and is not a directory; remove it first", mnt)
}

if err := patchWriteGuestFiles(ctx, disk); err != nil {
	if strings.Contains(err.Error(), "failed to create mount point") {
		log.WithError(err).Error("check permissions/ownership of the instance directory")
	}
	return err
}

When it happens

Trigger: Patch -> patchWriteGuestFiles when os.MkdirAll(mnt, 0o755) fails — the instance directory is not writable, a file named 'mnt' exists at that path, or the filesystem is read-only.

Common situations: LIMA_HOME on a read-only or restricted-permission volume; instance dir owned by another user; leftover file named 'mnt' from a previous crash; running limactl under a different user than the instance owner.

Understand the failure class

Background: mkdir permission denied (EACCES): failed to create directory errors explained — this error's family across 32 libraries.

Related errors

• failed to mount %#q on %#q: %w (output=%#q)
• failed to create temp file: %w
• failed to write temp plist: %w
• failed to create a temporary file for signing QEMU binary: %
• field `mountType` must be %#q or %#q for QEMU driver on macO

AI-assisted analysis of lima-vm/lima@dd909d0973 (2026-09-01). Data as JSON: /api/errors/8c526f08a30d6934. Report an issue: GitHub.