Do not reset storage when running inside of a container

Currently if the host shares container storage with a container
running podman, the podman inside of the container resets the
storage on the host. This can cause issues on the host, as
well as causes the podman command running the container, to
fail to unmount /dev/shm.

podman run -ti --rm --privileged -v /var/lib/containers:/var/lib/containers quay.io/podman/stable podman run alpine echo hello
	* unlinkat /var/lib/containers/storage/overlay-containers/a7f3c9deb0656f8de1d107e7ddff2d3c3c279c11c1635f233a0bffb16051fb2c/userdata/shm: device or resource busy
	* unlinkat /var/lib/containers/storage/overlay-containers/a7f3c9deb0656f8de1d107e7ddff2d3c3c279c11c1635f233a0bffb16051fb2c/userdata/shm: device or resource busy

Since podman is volume mounting in the graphroot, it will add a flag to
/run/.containerenv to tell podman inside of container whether to reset storage or not.

Since the inner podman is running inside of the container, no reason to assume this is a fresh reboot, so if "container" environment variable is set then skip
reset of storage.

Also added tests to make sure /run/.containerenv is runnig correctly.

Fixes: https://github.com/containers/podman/issues/9191

Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>

libpod/container_internal_linux.go

@@ -1627,7 +1627,7 @@ func (c *Container) makeBindMounts() error {
 
 	// Make .containerenv if it does not exist
 	if _, ok := c.state.BindMounts["/run/.containerenv"]; !ok {
-		var containerenv string
+		containerenv := c.runtime.graphRootMountedFlag(c.config.Spec.Mounts)
 		isRootless := 0
 		if rootless.IsRootless() {
 			isRootless = 1
@@ -1642,7 +1642,7 @@ id=%q
 image=%q
 imageid=%q
 rootless=%d
-`, version.Version.String(), c.Name(), c.ID(), imageName, imageID, isRootless)
+%s`, version.Version.String(), c.Name(), c.ID(), imageName, imageID, isRootless, containerenv)
 		}
 		containerenvPath, err := c.writeStringToRundir(".containerenv", containerenv)
 		if err != nil {

libpod/runtime.go

@@ -1,6 +1,7 @@
 package libpod
 
 import (
+	"bufio"
 	"context"
 	"fmt"
 	"os"
@@ -26,6 +27,7 @@ import (
 	"github.com/containers/storage"
 	"github.com/cri-o/ocicni/pkg/ocicni"
 	"github.com/docker/docker/pkg/namesgenerator"
+	spec "github.com/opencontainers/runtime-spec/specs-go"
 	"github.com/pkg/errors"
 	"github.com/sirupsen/logrus"
 )
@@ -622,9 +624,12 @@ func (r *Runtime) Shutdown(force bool) error {
 func (r *Runtime) refresh(alivePath string) error {
 	logrus.Debugf("Podman detected system restart - performing state refresh")
 
-	// First clear the state in the database
+	// Clear state of database if not running in container
-	if err := r.state.Refresh(); err != nil {
+	if !graphRootMounted() {
-		return err
+		// First clear the state in the database
+		if err := r.state.Refresh(); err != nil {
+			return err
+		}
 	}
 
 	// Next refresh the state of all containers to recreate dirs and
@@ -904,3 +909,29 @@ func (r *Runtime) getVolumePlugin(name string) (*plugin.VolumePlugin, error) {
 func (r *Runtime) GetSecretsStorageDir() string {
 	return filepath.Join(r.store.GraphRoot(), "secrets")
 }
+
+func graphRootMounted() bool {
+	f, err := os.OpenFile("/run/.containerenv", os.O_RDONLY, os.ModePerm)
+	if err != nil {
+		return false
+	}
+	defer f.Close()
+
+	scanner := bufio.NewScanner(f)
+	for scanner.Scan() {
+		if scanner.Text() == "graphRootMounted=1" {
+			return true
+		}
+	}
+	return false
+}
+
+func (r *Runtime) graphRootMountedFlag(mounts []spec.Mount) string {
+	root := r.store.GraphRoot()
+	for _, val := range mounts {
+		if strings.HasPrefix(root, val.Source) {
+			return "graphRootMounted=1"
+		}
+	}
+	return ""
+}

test/e2e/run_test.go

@@ -47,6 +47,29 @@ var _ = Describe("Podman run", func() {
 		Expect(session.ExitCode()).To(Equal(0))
 	})
 
+	It("podman run check /run/.containerenv", func() {
+		session := podmanTest.Podman([]string{"run", ALPINE, "cat", "/run/.containerenv"})
+		session.WaitWithDefaultTimeout()
+		Expect(session.ExitCode()).To(Equal(0))
+		Expect(session.OutputToString()).To(Equal(""))
+
+		session = podmanTest.Podman([]string{"run", "--privileged", "--name=test1", ALPINE, "cat", "/run/.containerenv"})
+		session.WaitWithDefaultTimeout()
+		Expect(session.ExitCode()).To(Equal(0))
+		Expect(session.OutputToString()).To(ContainSubstring("name=\"test1\""))
+		Expect(session.OutputToString()).To(ContainSubstring("image=\"" + ALPINE + "\""))
+
+		session = podmanTest.Podman([]string{"run", "-v", "/:/host", ALPINE, "cat", "/run/.containerenv"})
+		session.WaitWithDefaultTimeout()
+		Expect(session.ExitCode()).To(Equal(0))
+		Expect(session.OutputToString()).To(ContainSubstring("graphRootMounted=1"))
+
+		session = podmanTest.Podman([]string{"run", "-v", "/:/host", "--privileged", ALPINE, "cat", "/run/.containerenv"})
+		session.WaitWithDefaultTimeout()
+		Expect(session.ExitCode()).To(Equal(0))
+		Expect(session.OutputToString()).To(ContainSubstring("graphRootMounted=1"))
+	})
+
 	It("podman run a container based on a complex local image name", func() {
 		imageName := strings.TrimPrefix(nginx, "quay.io/")
 		session := podmanTest.Podman([]string{"run", imageName, "ls"})