Added option to keep containers running after checkpointing

CRIU supports to leave processes running after checkpointing:

  -R|--leave-running    leave tasks in running state after checkpoint

runc also support to leave containers running after checkpointing:

   --leave-running      leave the process running after checkpointing

With this commit the support to leave a container running after
checkpointing is brought to Podman:

   --leave-running, -R  leave the container running after writing checkpoint to disk

Now it is possible to checkpoint a container at some point in time
without stopping the container. This can be used to rollback the
container to an early state:

$ podman run --tmpfs /tmp --name podman-criu-test -d docker://docker.io/yovfiatbeb/podman-criu-test
$ curl 10.88.64.253:8080/examples/servlets/servlet/HelloWorldExample
3
$ podman container checkpoint -R -l
$ curl 10.88.64.253:8080/examples/servlets/servlet/HelloWorldExample
4
$ curl 10.88.64.253:8080/examples/servlets/servlet/HelloWorldExample
5
$ podman stop -l
$ podman container restore -l
$ curl 10.88.64.253:8080/examples/servlets/servlet/HelloWorldExample
4

So after checkpointing the container kept running and was stopped after
some time. Restoring this container will restore the state right at the
checkpoint.

Signed-off-by: Adrian Reber <areber@redhat.com>

4 files changed, 27 insertions, 10 deletions

diff --git a/cmd/podman/checkpoint.go b/cmd/podman/checkpoint.go
index 49e2aec63..ddfd12bc3 100644
--- a/cmd/podman/checkpoint.go
+++ b/cmd/podman/checkpoint.go
@@ -24,6 +24,10 @@ var (
 			Usage: "keep all temporary checkpoint files",
 		},
 		cli.BoolFlag{
+			Name:  "leave-running, R",
+			Usage: "leave the container running after writing checkpoint to disk",
+		},
+		cli.BoolFlag{
 			Name:  "all, a",
 			Usage: "checkpoint all running containers",
 		},
@@ -51,7 +55,8 @@ func checkpointCmd(c *cli.Context) error {
 	defer runtime.Shutdown(false)
 
 	options := libpod.ContainerCheckpointOptions{
-		Keep: c.Bool("keep"),
+		Keep:        c.Bool("keep"),
+		KeepRunning: c.Bool("leave-running"),
 	}
 
 	if err := checkAllAndLatest(c); err != nil {
diff --git a/libpod/container_api.go b/libpod/container_api.go
index 9f5436b14..df6b6e962 100644
--- a/libpod/container_api.go
+++ b/libpod/container_api.go
@@ -833,7 +833,8 @@ func (c *Container) Refresh(ctx context.Context) error {
 // ContainerCheckpointOptions is a struct used to pass the parameters
 // for checkpointing to corresponding functions
 type ContainerCheckpointOptions struct {
-	Keep bool
+	Keep        bool
+	KeepRunning bool
 }
 
 // Checkpoint checkpoints a container
diff --git a/libpod/container_internal_linux.go b/libpod/container_internal_linux.go
index 003e8284a..e6071945d 100644
--- a/libpod/container_internal_linux.go
+++ b/libpod/container_internal_linux.go
@@ -440,7 +440,7 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 	if c.state.State != ContainerStateRunning {
 		return errors.Wrapf(ErrCtrStateInvalid, "%q is not running, cannot checkpoint", c.state.State)
 	}
-	if err := c.runtime.ociRuntime.checkpointContainer(c); err != nil {
+	if err := c.runtime.ociRuntime.checkpointContainer(c, options); err != nil {
 		return err
 	}
 
@@ -457,11 +457,13 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 
 	logrus.Debugf("Checkpointed container %s", c.ID())
 
-	c.state.State = ContainerStateStopped
+	if !options.KeepRunning {
+		c.state.State = ContainerStateStopped
 
-	// Cleanup Storage and Network
-	if err := c.cleanup(ctx); err != nil {
-		return err
+		// Cleanup Storage and Network
+		if err := c.cleanup(ctx); err != nil {
+			return err
+		}
 	}
 
 	if !options.Keep {
diff --git a/libpod/oci.go b/libpod/oci.go
index 71da830b5..8ee2c948f 100644
--- a/libpod/oci.go
+++ b/libpod/oci.go
@@ -844,13 +844,22 @@ func (r *OCIRuntime) execStopContainer(ctr *Container, timeout uint) error {
 }
 
 // checkpointContainer checkpoints the given container
-func (r *OCIRuntime) checkpointContainer(ctr *Container) error {
+func (r *OCIRuntime) checkpointContainer(ctr *Container, options ContainerCheckpointOptions) error {
 	// imagePath is used by CRIU to store the actual checkpoint files
 	imagePath := ctr.CheckpointPath()
 	// workPath will be used to store dump.log and stats-dump
 	workPath := ctr.bundlePath()
 	logrus.Debugf("Writing checkpoint to %s", imagePath)
 	logrus.Debugf("Writing checkpoint logs to %s", workPath)
-	return utils.ExecCmdWithStdStreams(os.Stdin, os.Stdout, os.Stderr, nil, r.path, "checkpoint",
-		"--image-path", imagePath, "--work-path", workPath, ctr.ID())
+	args := []string{}
+	args = append(args, "checkpoint")
+	args = append(args, "--image-path")
+	args = append(args, imagePath)
+	args = append(args, "--work-path")
+	args = append(args, workPath)
+	if options.KeepRunning {
+		args = append(args, "--leave-running")
+	}
+	args = append(args, ctr.ID())
+	return utils.ExecCmdWithStdStreams(os.Stdin, os.Stdout, os.Stderr, nil, r.path, args...)
 }