kubernetes/kops · error

error checking for /run/cilium/cgroupv2: %v

Error message

error checking for /run/cilium/cgroupv2: %v

What it means

buildCgroup2Mount calls unix.Statfs on /run/cilium/cgroupv2; if the error is anything other than os.ErrNotExist (which is tolerated because systemd creates the path), the error is wrapped and the build fails. This guards against unexpected filesystem errors on the cgroup path.

Source

Thrown at nodeup/pkg/model/networking/cilium.go:124

			Name:       "sys-fs-bpf.mount",
			Definition: new(unit),
		}
		service.InitDefaults()
		c.AddTask(service)
	}

	return nil
}

func (b *CiliumBuilder) buildCgroup2Mount(c *fi.NodeupModelBuilderContext) error {
	cgroupPath := "/run/cilium/cgroupv2"

	var fsdata unix.Statfs_t
	err := unix.Statfs(cgroupPath, &fsdata)

	// If the path does not exist, systemd will create it
	if err != nil && !errors.Is(err, os.ErrNotExist) {
		return fmt.Errorf("error checking for /run/cilium/cgroupv2: %v", err)
	}

	CGROUP_FS_MAGIC := uint32(0x63677270)

	alreadyMounted := uint32(fsdata.Type) == CGROUP_FS_MAGIC

	if !alreadyMounted {
		unit := `
[Unit]
Description=Cilium Cgroup2 mounts
Documentation=http://docs.cilium.io/
DefaultDependencies=no
Before=local-fs.target umount.target kubelet.service

[Mount]
What=cgroup2
Where=/run/cilium/cgroupv2
Type=cgroup2

View on GitHub (pinned to 4c8573c808)

Solutions

1. Check permissions and type of /run/cilium/cgroupv2 (remove a stray non-directory file: rm and let systemd recreate it)
2. Ensure /run is a writable tmpfs mount when nodeup runs
3. Read the wrapped errno in the message and fix the underlying filesystem condition
4. Reboot/clean stale state in /run/cilium and re-run nodeup

Example fix

// before
$ ls -ld /run/cilium/cgroupv2
-rw-r--r-- 1 root root 0 ...   # stray file, not a directory
// after
$ rm /run/cilium/cgroupv2   # systemd will create the cgroup dir

var fsdata unix.Statfs_t
err := unix.Statfs("/run/cilium/cgroupv2", &fsdata)
if err != nil && !errors.Is(err, os.ErrNotExist) {
    // inspect errno
    klog.Errorf("unexpected cgroupv2 stat error: %v", err)
}
if fi, statErr := os.Stat("/run/cilium/cgroupv2"); statErr == nil && !fi.IsDir() {
    klog.Warning("/run/cilium/cgroupv2 is not a directory; remove it")
}

err := unix.Statfs(cgroupPath, &fsdata)
if err != nil && !errors.Is(err, os.ErrNotExist) {
    return fmt.Errorf("error checking for /run/cilium/cgroupv2: %v", err)
}

When it happens

Trigger: unix.Statfs(cgroupPath) fails with e.g. EACCES, ENOTDIR, ELOOP, or an I/O error — anything besides ENOENT — during nodeup Build with Cilium enabled.

Common situations: /run not mounted or read-only; a non-directory file occupying /run/cilium/cgroupv2; permission/hardening policies blocking stat on /run paths; symlink loops from earlier failed runs.

Related errors

• failed to create cgroupv2 mount unit: %w
• error checking for /sys/fs/bpf: %v
• path: %s already exists but is not a directory
• failed to read %s: %w
• failed to create bpf mount unit: %w

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/57006d2ffec0237d. Report an issue: GitHub.