Add ContainerStateRemoving

When Libpod removes a container, there is the possibility that
removal will not fully succeed. The most notable problems are
storage issues, where the container cannot be removed from
c/storage.

When this occurs, we were faced with a choice. We can keep the
container in the state, appearing in `podman ps` and available for
other API operations, but likely unable to do any of them as it's
been partially removed. Or we can remove it very early and clean
up after it's already gone. We have, until now, used the second
approach.

The problem that arises is intermittent problems removing
storage. We end up removing a container, failing to remove its
storage, and ending up with a container permanently stuck in
c/storage that we can't remove with the normal Podman CLI, can't
use the name of, and generally can't interact with. A notable
cause is when Podman is hit by a SIGKILL midway through removal,
which can consistently cause `podman rm` to fail to remove
storage.

We now add a new state for containers that are in the process of
being removed, ContainerStateRemoving. We set this at the
beginning of the removal process. It notifies Podman that the
container cannot be used anymore, but preserves it in the DB
until it is fully removed. This will allow Remove to be run on
these containers again, which should successfully remove storage
if it fails.

Fixes containers#3906

Signed-off-by: Matthew Heon <[email protected]>

cmd/podman/shared/container.go

@@ -195,6 +195,8 @@ func NewBatchContainer(ctr *libpod.Container, opts PsOptions) (PsContainerOutput
 		status = "Paused"
 	case define.ContainerStateCreated.String(), define.ContainerStateConfigured.String():
 		status = "Created"
+	case define.ContainerStateRemoving.String():
+		status = "Removing"
 	default:
 		status = "Error"
 	}

libpod/container_api.go

@@ -404,6 +404,11 @@ func (c *Container) Mount() (string, error) {
 			return "", err
 		}
 	}
+
+	if c.state.State == define.ContainerStateRemoving {
+		return "", errors.Wrapf(define.ErrCtrStateInvalid, "cannot mount container %s as it is being removed", c.ID())
+	}
+
 	defer c.newContainerEvent(events.Mount)
 	return c.mount()
 }
@@ -488,7 +493,12 @@ func (c *Container) Export(path string) error {
 			return err
 		}
 	}
-	defer c.newContainerEvent(events.Export)
+
+	if c.state.State == define.ContainerStateRemoving {
+		return errors.Wrapf(define.ErrCtrStateInvalid, "cannot mount container %s as it is being removed", c.ID())
+	}
+
+	defer c.newContainerEvent(events.Mount)
 	return c.export(path)
 }
 
@@ -674,6 +684,10 @@ func (c *Container) Refresh(ctx context.Context) error {
 		}
 	}
 
+	if c.state.State == define.ContainerStateRemoving {
+		return errors.Wrapf(define.ErrCtrStateInvalid, "cannot refresh containers that are being removed")
+	}
+
 	wasCreated := false
 	if c.state.State == define.ContainerStateCreated {
 		wasCreated = true
@@ -819,7 +833,6 @@ func (c *Container) Checkpoint(ctx context.Context, options ContainerCheckpointO
 			return err
 		}
 	}
-	defer c.newContainerEvent(events.Checkpoint)
 	return c.checkpoint(ctx, options)
 }
 

libpod/container_internal.go

@@ -719,7 +719,8 @@ func (c *Container) isStopped() (bool, error) {
 	if err != nil {
 		return true, err
 	}
-	return c.state.State != define.ContainerStateRunning && c.state.State != define.ContainerStatePaused, nil
+
+	return !c.ensureState(define.ContainerStateRunning, define.ContainerStatePaused), nil
 }
 
 // save container state to the database
@@ -1057,6 +1058,8 @@ func (c *Container) initAndStart(ctx context.Context) (err error) {
 	// If we are ContainerStateUnknown, throw an error
 	if c.state.State == define.ContainerStateUnknown {
 		return errors.Wrapf(define.ErrCtrStateInvalid, "container %s is in an unknown state", c.ID())
+	} else if c.state.State == define.ContainerStateRemoving {
+		return errors.Wrapf(define.ErrCtrStateInvalid, "cannot start container %s as it is being removed", c.ID())
 	}
 
 	// If we are running, do nothing

libpod/container_internal_linux.go

@@ -21,6 +21,7 @@ import (
 	"github.com/containernetworking/plugins/pkg/ns"
 	"github.com/containers/buildah/pkg/secrets"
 	"github.com/containers/libpod/libpod/define"
+	"github.com/containers/libpod/libpod/events"
 	"github.com/containers/libpod/pkg/annotations"
 	"github.com/containers/libpod/pkg/apparmor"
 	"github.com/containers/libpod/pkg/cgroups"
@@ -695,6 +696,8 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 		return err
 	}
 
+	defer c.newContainerEvent(events.Checkpoint)
+
 	if options.TargetFile != "" {
 		if err = c.exportCheckpoint(options.TargetFile, options.IgnoreRootfs); err != nil {
 			return err
@@ -766,7 +769,7 @@ func (c *Container) restore(ctx context.Context, options ContainerCheckpointOpti
 		return err
 	}
 
-	if (c.state.State != define.ContainerStateConfigured) && (c.state.State != define.ContainerStateExited) {
+	if !c.ensureState(define.ContainerStateConfigured, define.ContainerStateExited) {
 		return errors.Wrapf(define.ErrCtrStateInvalid, "container %s is running or paused, cannot restore", c.ID())
 	}
 

libpod/define/containerstate.go

@@ -25,6 +25,9 @@ const (
 	// ContainerStateExited indicates the the container has stopped and been
 	// cleaned up
 	ContainerStateExited ContainerStatus = iota
+	// ContainerStateRemoving indicates the container is in the process of
+	// being removed.
+	ContainerStateRemoving ContainerStatus = iota
 )
 
 // ContainerStatus returns a string representation for users
@@ -45,6 +48,8 @@ func (t ContainerStatus) String() string {
 		return "paused"
 	case ContainerStateExited:
 		return "exited"
+	case ContainerStateRemoving:
+		return "removing"
 	}
 	return "bad state"
 }
@@ -67,6 +72,8 @@ func StringToContainerStatus(status string) (ContainerStatus, error) {
 		return ContainerStatePaused, nil
 	case ContainerStateExited.String():
 		return ContainerStateExited, nil
+	case ContainerStateRemoving.String():
+		return ContainerStateRemoving, nil
 	default:
 		return ContainerStateUnknown, errors.Wrapf(ErrInvalidArg, "unknown container state: %s", status)
 	}

libpod/options.go

@@ -768,16 +768,8 @@ func WithIPCNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
-		}
-
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		ctr.config.IPCNsCtr = nsCtr.ID()
@@ -796,16 +788,8 @@ func WithMountNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
-		}
-
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		ctr.config.MountNsCtr = nsCtr.ID()
@@ -824,22 +808,14 @@ func WithNetNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		if ctr.config.CreateNetNS {
 			return errors.Wrapf(define.ErrInvalidArg, "cannot join another container's net ns as we are making a new net ns")
 		}
 
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
-		}
-
 		ctr.config.NetNsCtr = nsCtr.ID()
 
 		return nil
@@ -856,16 +832,8 @@ func WithPIDNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
-		}
-
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		if ctr.config.NoCgroups {
@@ -888,16 +856,8 @@ func WithUserNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
-		}
-
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		ctr.config.UserNsCtr = nsCtr.ID()
@@ -917,16 +877,8 @@ func WithUTSNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
-		}
-
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		ctr.config.UTSNsCtr = nsCtr.ID()
@@ -945,16 +897,8 @@ func WithCgroupNSFrom(nsCtr *Container) CtrCreateOption {
 			return define.ErrCtrFinalized
 		}
 
-		if !nsCtr.valid {
-			return define.ErrCtrRemoved
-		}
-
-		if nsCtr.ID() == ctr.ID() {
-			return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
-		}
-
-		if ctr.config.Pod != "" && nsCtr.config.Pod != ctr.config.Pod {
-			return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, nsCtr.ID())
+		if err := checkDependencyContainer(nsCtr, ctr); err != nil {
+			return err
 		}
 
 		ctr.config.CgroupNsCtr = nsCtr.ID()

libpod/runtime_ctr.go

@@ -489,32 +489,19 @@ func (r *Runtime) removeContainer(ctx context.Context, c *Container, force bool,
 		}
 	}
 
-	var cleanupErr error
-	// Remove the container from the state
-	if c.config.Pod != "" {
-		// If we're removing the pod, the container will be evicted
-		// from the state elsewhere
-		if !removePod {
-			if err := r.state.RemoveContainerFromPod(pod, c); err != nil {
-				cleanupErr = err
-			}
-		}
-	} else {
-		if err := r.state.RemoveContainer(c); err != nil {
-			cleanupErr = err
-		}
+	// Set ContainerStateRemoving and remove exec sessions
+	c.state.State = define.ContainerStateRemoving
+	c.state.ExecSessions = nil
+
+	if err := c.save(); err != nil {
+		return errors.Wrapf(err, "unable to set container %s removing state in database", c.ID())
 	}
 
-	// Set container as invalid so it can no longer be used
-	c.valid = false
+	var cleanupErr error
 
 	// Clean up network namespace, cgroups, mounts
 	if err := c.cleanup(ctx); err != nil {
-		if cleanupErr == nil {
-			cleanupErr = errors.Wrapf(err, "error cleaning up container %s", c.ID())
-		} else {
-			logrus.Errorf("cleanup network, cgroups, mounts: %v", err)
-		}
+		cleanupErr = errors.Wrapf(err, "error cleaning up container %s", c.ID())
 	}
 
 	// Stop the container's storage
@@ -540,6 +527,29 @@ func (r *Runtime) removeContainer(ctx context.Context, c *Container, force bool,
 		}
 	}
 
+	// Remove the container from the state
+	if c.config.Pod != "" {
+		// If we're removing the pod, the container will be evicted
+		// from the state elsewhere
+		if !removePod {
+			if err := r.state.RemoveContainerFromPod(pod, c); err != nil {
+				if cleanupErr == nil {
+					cleanupErr = err
+				} else {
+					logrus.Errorf("Error removing container %s from database: %v", c.ID(), err)
+				}
+			}
+		}
+	} else {
+		if err := r.state.RemoveContainer(c); err != nil {
+			if cleanupErr == nil {
+				cleanupErr = err
+			} else {
+				logrus.Errorf("Error removing container %s from database: %v", c.ID(), err)
+			}
+		}
+	}
+
 	// Deallocate the container's lock
 	if err := c.lock.Free(); err != nil {
 		if cleanupErr == nil {
@@ -549,6 +559,9 @@ func (r *Runtime) removeContainer(ctx context.Context, c *Container, force bool,
 		}
 	}
 
+	// Set container as invalid so it can no longer be used
+	c.valid = false
+
 	c.newContainerEvent(events.Remove)
 
 	if !removeVolume {

libpod/util.go

@@ -206,3 +206,28 @@ func DefaultSeccompPath() (string, error) {
 	}
 	return config.SeccompDefaultPath, nil
 }
+
+// CheckDependencyContainer verifies the given container can be used as a
+// dependency of another container.
+// Both the dependency to check and the container that will be using the
+// dependency must be passed in.
+// It is assumed that ctr is locked, and depCtr is unlocked.
+func checkDependencyContainer(depCtr, ctr *Container) error {
+	state, err := depCtr.State()
+	if err != nil {
+		return errors.Wrapf(err, "error accessing dependency container %s state", depCtr.ID())
+	}
+	if state == define.ContainerStateRemoving {
+		return errors.Wrapf(define.ErrCtrStateInvalid, "cannot use container %s as a dependency as it is being removed", depCtr.ID())
+	}
+
+	if depCtr.ID() == ctr.ID() {
+		return errors.Wrapf(define.ErrInvalidArg, "must specify another container")
+	}
+
+	if ctr.config.Pod != "" && depCtr.PodID() != ctr.config.Pod {
+		return errors.Wrapf(define.ErrInvalidArg, "container has joined pod %s and dependency container %s is not a member of the pod", ctr.config.Pod, depCtr.ID())
+	}
+
+	return nil
+}