Added optional container checkpointing statistics

This adds the parameter '--print-stats' to 'podman container checkpoint'.
With '--print-stats' Podman will measure how long Podman itself, the OCI
runtime and CRIU requires to create a checkpoint and print out these
information. CRIU already creates checkpointing statistics which are
just read in addition to the added measurements. In contrast to just
printing out the ID of the checkpointed container, Podman will now print
out JSON:

 # podman container checkpoint --latest --print-stats
 {
     "podman_checkpoint_duration": 360749,
     "container_statistics": [
         {
             "Id": "25244244bf2efbef30fb6857ddea8cb2e5489f07eb6659e20dda117f0c466808",
             "runtime_checkpoint_duration": 177222,
             "criu_statistics": {
                 "freezing_time": 100657,
                 "frozen_time": 60700,
                 "memdump_time": 8162,
                 "memwrite_time": 4224,
                 "pages_scanned": 20561,
                 "pages_written": 2129
             }
         }
     ]
 }

The output contains 'podman_checkpoint_duration' which contains the
number of microseconds Podman required to create the checkpoint. The
output also includes 'runtime_checkpoint_duration' which is the time
the runtime needed to checkpoint that specific container. Each container
also includes 'criu_statistics' which displays the timing information
collected by CRIU.

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

1 files changed, 40 insertions, 12 deletions

diff --git a/libpod/container_internal_linux.go b/libpod/container_internal_linux.go
index 91453574e..2778d2cde 100644
--- a/libpod/container_internal_linux.go
+++ b/libpod/container_internal_linux.go
@@ -1129,25 +1129,26 @@ func (c *Container) checkpointRestoreSupported(version int) error {
 	return nil
 }
 
-func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointOptions) error {
+func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointOptions) (*define.CRIUCheckpointRestoreStatistics, int64, error) {
 	if err := c.checkpointRestoreSupported(criu.MinCriuVersion); err != nil {
-		return err
+		return nil, 0, err
 	}
 
 	if c.state.State != define.ContainerStateRunning {
-		return errors.Wrapf(define.ErrCtrStateInvalid, "%q is not running, cannot checkpoint", c.state.State)
+		return nil, 0, errors.Wrapf(define.ErrCtrStateInvalid, "%q is not running, cannot checkpoint", c.state.State)
 	}
 
 	if c.AutoRemove() && options.TargetFile == "" {
-		return errors.Errorf("cannot checkpoint containers that have been started with '--rm' unless '--export' is used")
+		return nil, 0, errors.Errorf("cannot checkpoint containers that have been started with '--rm' unless '--export' is used")
 	}
 
 	if err := crutils.CRCreateFileWithLabel(c.bundlePath(), "dump.log", c.MountLabel()); err != nil {
-		return err
+		return nil, 0, err
 	}
 
-	if err := c.ociRuntime.CheckpointContainer(c, options); err != nil {
-		return err
+	runtimeCheckpointDuration, err := c.ociRuntime.CheckpointContainer(c, options)
+	if err != nil {
+		return nil, 0, err
 	}
 
 	// Save network.status. This is needed to restore the container with
@@ -1155,7 +1156,7 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 	// with one interface.
 	// FIXME: will this break something?
 	if _, err := metadata.WriteJSONFile(c.getNetworkStatus(), c.bundlePath(), metadata.NetworkStatusFile); err != nil {
-		return err
+		return nil, 0, err
 	}
 
 	defer c.newContainerEvent(events.Checkpoint)
@@ -1165,13 +1166,13 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 	if options.WithPrevious {
 		os.Remove(path.Join(c.CheckpointPath(), "parent"))
 		if err := os.Symlink("../pre-checkpoint", path.Join(c.CheckpointPath(), "parent")); err != nil {
-			return err
+			return nil, 0, err
 		}
 	}
 
 	if options.TargetFile != "" {
 		if err := c.exportCheckpoint(options); err != nil {
-			return err
+			return nil, 0, err
 		}
 	}
 
@@ -1183,8 +1184,35 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 
 		// Cleanup Storage and Network
 		if err := c.cleanup(ctx); err != nil {
-			return err
+			return nil, 0, err
+		}
+	}
+
+	criuStatistics, err := func() (*define.CRIUCheckpointRestoreStatistics, error) {
+		if !options.PrintStats {
+			return nil, nil
+		}
+		statsDirectory, err := os.Open(c.bundlePath())
+		if err != nil {
+			return nil, errors.Wrapf(err, "Not able to open %q", c.bundlePath())
+		}
+
+		dumpStatistics, err := stats.CriuGetDumpStats(statsDirectory)
+		if err != nil {
+			return nil, errors.Wrap(err, "Displaying checkpointing statistics not possible")
 		}
+
+		return &define.CRIUCheckpointRestoreStatistics{
+			FreezingTime: dumpStatistics.GetFreezingTime(),
+			FrozenTime:   dumpStatistics.GetFrozenTime(),
+			MemdumpTime:  dumpStatistics.GetMemdumpTime(),
+			MemwriteTime: dumpStatistics.GetMemwriteTime(),
+			PagesScanned: dumpStatistics.GetPagesScanned(),
+			PagesWritten: dumpStatistics.GetPagesWritten(),
+		}, nil
+	}()
+	if err != nil {
+		return nil, 0, err
 	}
 
 	if !options.Keep && !options.PreCheckPoint {
@@ -1203,7 +1231,7 @@ func (c *Container) checkpoint(ctx context.Context, options ContainerCheckpointO
 	}
 
 	c.state.FinishedTime = time.Now()
-	return c.save()
+	return criuStatistics, runtimeCheckpointDuration, c.save()
 }
 
 func (c *Container) importCheckpoint(input string) error {