Fix deletion of data and kube root dirs

Make sure that we don't have anything mounted under those directories so
we don't delete persistent data.

We do this py parsing /proc/mounts in reverse order as it is listed in
mount order, and then we unmount anything that is under our directories
before we delete them.

Don't umount datadir itself if it is on a separate partition/mount

Fixes #4318

Signed-off-by: Natanael Copa <[email protected]>

pkg/cleanup/directories.go

@@ -21,8 +21,10 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"strings"
 
 	"github.com/sirupsen/logrus"
+	"golang.org/x/sys/unix"
 	"k8s.io/mount-utils"
 )
 
@@ -46,15 +48,23 @@ func (d *directories) Run() error {
 
 	var dataDirMounted bool
 
-	// search and unmount kubelet volume mounts
-	for _, v := range procMounts {
-		if v.Path == filepath.Join(d.Config.dataDir, "kubelet") {
+	// search and unmount anything under kubelet root or data dir
+	// in reverse order to handle nested mounts and over mounts
+	for i := len(procMounts) - 1; i >= 0; i-- {
+		v := procMounts[i]
+		// avoid unmount datadir if its mounted on separate partition
+		if v.Path == d.Config.k0sVars.DataDir {
+			dataDirMounted = true
+			continue
+		}
+		if isUnderPath(v.Path, filepath.Join(d.Config.dataDir, "kubelet")) || isUnderPath(v.Path, d.Config.k0sVars.DataDir) {
 			logrus.Debugf("%v is mounted! attempting to unmount...", v.Path)
 			if err = mounter.Unmount(v.Path); err != nil {
-				logrus.Warningf("failed to unmount %v", v.Path)
+				logrus.Warningf("lazy unmounting %v", v.Path)
+				if err = unix.Unmount(v.Path, unix.MNT_DETACH); err != nil {
+					return fmt.Errorf("failed unmount %v", v.Path)
+				}
 			}
-		} else if v.Path == d.Config.dataDir {
-			dataDirMounted = true
 		}
 	}
 
@@ -81,6 +91,12 @@ func (d *directories) Run() error {
 	return nil
 }
 
+// test if the path is a directory equal to or under base
+func isUnderPath(path, base string) bool {
+	rel, err := filepath.Rel(base, path)
+	return err == nil && !strings.HasPrefix(rel, "..") && !filepath.IsAbs(rel)
+}
+
 // this is for checking if the error retrned by os.RemoveAll is due to
 // it being a mount point. if it is, we can ignore the error. this way
 // we can't rely on os.RemoveAll instead of recursively deleting the