keep a consistent view of containers rendered

Replicate relevant mutations to the in-memory ACID store. Readers will
then be able to query container state without locking.

Signed-off-by: Fabio Kung <fabio.kung@gmail.com>

container/container_unix.go

@@ -262,11 +262,8 @@ func (container *Container) ConfigMounts() []Mount {
 	return mounts
 }
 
-// UpdateContainer updates configuration of a container.
+// UpdateContainer updates configuration of a container. Callers must hold a Lock on the Container.
 func (container *Container) UpdateContainer(hostConfig *containertypes.HostConfig) error {
-	container.Lock()
-	defer container.Unlock()
-
 	// update resources of container
 	resources := hostConfig.Resources
 	cResources := &container.HostConfig.Resources

container/container_windows.go

@@ -126,11 +126,8 @@ func (container *Container) TmpfsMounts() ([]Mount, error) {
 	return mounts, nil
 }
 
-// UpdateContainer updates configuration of a container
+// UpdateContainer updates configuration of a container. Callers must hold a Lock on the Container.
 func (container *Container) UpdateContainer(hostConfig *containertypes.HostConfig) error {
-	container.Lock()
-	defer container.Unlock()
-
 	resources := hostConfig.Resources
 	if resources.CPUShares != 0 ||
 		resources.Memory != 0 ||

daemon/container.go

@@ -99,7 +99,7 @@ func (daemon *Daemon) load(id string) (*container.Container, error) {
 }
 
 // Register makes a container object usable by the daemon as <container.ID>
-func (daemon *Daemon) Register(c *container.Container) {
+func (daemon *Daemon) Register(c *container.Container) error {
 	// Attach to stdout and stderr
 	if c.Config.OpenStdin {
 		c.StreamConfig.NewInputPipes()
@@ -107,8 +107,14 @@ func (daemon *Daemon) Register(c *container.Container) {
 		c.StreamConfig.NewNopInputPipe()
 	}
 
+	// once in the memory store it is visible to other goroutines
+	// grab a Lock until it has been replicated to avoid races
+	c.Lock()
+	defer c.Unlock()
+
 	daemon.containers.Add(c.ID, c)
 	daemon.idIndex.Add(c.ID)
+	return daemon.containersReplica.Save(c.Snapshot())
 }
 
 func (daemon *Daemon) newContainer(name string, platform string, config *containertypes.Config, hostConfig *containertypes.HostConfig, imgID image.ID, managed bool) (*container.Container, error) {
@@ -212,6 +218,9 @@ func (daemon *Daemon) setHostConfig(container *container.Container, hostConfig *
 
 	runconfig.SetDefaultNetModeIfBlank(hostConfig)
 	container.HostConfig = hostConfig
+	if err := daemon.containersReplica.Save(container.Snapshot()); err != nil {
+		return err
+	}
 	return container.ToDisk()
 }
 

daemon/container_operations.go

@@ -44,6 +44,19 @@ func (daemon *Daemon) getDNSSearchSettings(container *container.Container) []str
 
 	return nil
 }
+
+func (daemon *Daemon) saveAndReplicate(container *container.Container) error {
+	container.Lock()
+	defer container.Unlock()
+	if err := daemon.containersReplica.Save(container.Snapshot()); err != nil {
+		return fmt.Errorf("Error replicating container state: %v", err)
+	}
+	if err := container.ToDisk(); err != nil {
+		return fmt.Errorf("Error saving container to disk: %v", err)
+	}
+	return nil
+}
+
 func (daemon *Daemon) buildSandboxOptions(container *container.Container) ([]libnetwork.SandboxOption, error) {
 	var (
 		sboxOptions []libnetwork.SandboxOption
@@ -1005,7 +1018,7 @@ func (daemon *Daemon) ConnectToNetwork(container *container.Container, idOrName
 			return err
 		}
 	}
-	if err := container.ToDisk(); err != nil {
+	if err := daemon.saveAndReplicate(container); err != nil {
 		return fmt.Errorf("Error saving container to disk: %v", err)
 	}
 	return nil
@@ -1044,16 +1057,16 @@ func (daemon *Daemon) DisconnectFromNetwork(container *container.Container, netw
 		return err
 	}
 
-	if err := container.ToDisk(); err != nil {
+	if err := daemon.saveAndReplicate(container); err != nil {
 		return fmt.Errorf("Error saving container to disk: %v", err)
 	}
 
 	if n != nil {
-		attributes := map[string]string{
+		daemon.LogNetworkEventWithAttributes(n, "disconnect", map[string]string{
 			"container": container.ID,
+		})
 	}
-		daemon.LogNetworkEventWithAttributes(n, "disconnect", attributes)
-	}
+
 	return nil
 }
 

daemon/create.go

@@ -172,7 +172,9 @@ func (daemon *Daemon) create(params types.ContainerCreateConfig, managed bool) (
 		logrus.Errorf("Error saving new container to disk: %v", err)
 		return nil, err
 	}
-	daemon.Register(container)
+	if err := daemon.Register(container); err != nil {
+		return nil, err
+	}
 	stateCtr.set(container.ID, "stopped")
 	daemon.LogContainerEvent(container, "create")
 	return container, nil

daemon/daemon.go

@@ -83,6 +83,7 @@ type Daemon struct {
 	ID                    string
 	repository            string
 	containers            container.Store
+	containersReplica     *container.MemDB
 	execCommands          *exec.Store
 	downloadManager       *xfer.LayerDownloadManager
 	uploadManager         *xfer.LayerUploadManager
@@ -182,11 +183,15 @@ func (daemon *Daemon) restore() error {
 	activeSandboxes := make(map[string]interface{})
 	for id, c := range containers {
 		if err := daemon.registerName(c); err != nil {
+			logrus.Errorf("Failed to register container name %s: %s", c.ID, err)
+			delete(containers, id)
+			continue
+		}
+		if err := daemon.Register(c); err != nil {
 			logrus.Errorf("Failed to register container %s: %s", c.ID, err)
 			delete(containers, id)
 			continue
 		}
-		daemon.Register(c)
 
 		// verify that all volumes valid and have been migrated from the pre-1.7 layout
 		if err := daemon.verifyVolumesInfo(c); err != nil {
@@ -757,6 +762,9 @@ func NewDaemon(config *config.Config, registryService registry.Service, containe
 	d.ID = trustKey.PublicKey().KeyID()
 	d.repository = daemonRepo
 	d.containers = container.NewMemoryStore()
+	if d.containersReplica, err = container.NewMemDB(); err != nil {
+		return nil, err
+	}
 	d.execCommands = exec.NewStore()
 	d.trustKey = trustKey
 	d.idIndex = truncindex.NewTruncIndex([]string{})

daemon/delete.go

@@ -103,14 +103,20 @@ func (daemon *Daemon) cleanupContainer(container *container.Container, forceRemo
 	}
 
 	// Mark container dead. We don't want anybody to be restarting it.
-	container.SetDead()
+	container.Lock()
+	container.Dead = true
+	if err = daemon.containersReplica.Save(container.Snapshot()); err != nil {
+		container.Unlock()
+		return err
+	}
 
 	// Save container state to disk. So that if error happens before
 	// container meta file got removed from disk, then a restart of
 	// docker should not make a dead container alive.
-	if err := container.ToDiskLocking(); err != nil && !os.IsNotExist(err) {
+	if err := container.ToDisk(); err != nil && !os.IsNotExist(err) {
 		logrus.Errorf("Error saving dying container to disk: %v", err)
 	}
+	container.Unlock()
 
 	// When container creation fails and `RWLayer` has not been created yet, we
 	// do not call `ReleaseRWLayer`
@@ -131,6 +137,7 @@ func (daemon *Daemon) cleanupContainer(container *container.Container, forceRemo
 	selinuxFreeLxcContexts(container.ProcessLabel)
 	daemon.idIndex.Delete(container.ID)
 	daemon.containers.Delete(container.ID)
+	daemon.containersReplica.Delete(container.ID)
 	if e := daemon.removeMountPoints(container, removeVolume); e != nil {
 		logrus.Error(e)
 	}

daemon/health.go

@@ -167,6 +167,13 @@ func handleProbeResult(d *Daemon, c *container.Container, result *types.Healthch
 		// Else we're starting or healthy. Stay in that state.
 	}
 
+	// replicate Health status changes
+	if err := d.containersReplica.Save(c.Snapshot()); err != nil {
+		// queries will be inconsistent until the next probe runs or other state mutations
+		// trigger a replication
+		logrus.Errorf("Error replicating health state for container %s: %v", c.ID, err)
+	}
+
 	if oldStatus != h.Status {
 		d.LogContainerEvent(c, "health_status: "+h.Status)
 	}

daemon/health_test.go

@@ -29,7 +29,13 @@ func TestNoneHealthcheck(t *testing.T) {
 		},
 		State: &container.State{},
 	}
-	daemon := &Daemon{}
+	store, err := container.NewMemDB()
+	if err != nil {
+		t.Fatal(err)
+	}
+	daemon := &Daemon{
+		containersReplica: store,
+	}
 
 	daemon.initHealthMonitor(c)
 	if c.State.Health != nil {
@@ -62,8 +68,15 @@ func TestHealthStates(t *testing.T) {
 			Image: "image_name",
 		},
 	}
+
+	store, err := container.NewMemDB()
+	if err != nil {
+		t.Fatal(err)
+	}
+
 	daemon := &Daemon{
 		EventsService:     e,
+		containersReplica: store,
 	}
 
 	c.Config.Healthcheck = &containertypes.HealthConfig{

daemon/monitor.go

@@ -90,6 +90,9 @@ func (daemon *Daemon) StateChanged(id string, e libcontainerd.StateInfo) error {
 		daemon.setStateCounter(c)
 
 		defer c.Unlock()
+		if err := daemon.containersReplica.Save(c.Snapshot()); err != nil {
+			return err
+		}
 		if err := c.ToDisk(); err != nil {
 			return err
 		}
@@ -119,6 +122,10 @@ func (daemon *Daemon) StateChanged(id string, e libcontainerd.StateInfo) error {
 		c.HasBeenStartedBefore = true
 		daemon.setStateCounter(c)
 
+		if err := daemon.containersReplica.Save(c.Snapshot()); err != nil {
+			c.Reset(false)
+			return err
+		}
 		if err := c.ToDisk(); err != nil {
 			c.Reset(false)
 			return err
@@ -130,6 +137,9 @@ func (daemon *Daemon) StateChanged(id string, e libcontainerd.StateInfo) error {
 		// Container is already locked in this case
 		c.Paused = true
 		daemon.setStateCounter(c)
+		if err := daemon.containersReplica.Save(c.Snapshot()); err != nil {
+			return err
+		}
 		if err := c.ToDisk(); err != nil {
 			return err
 		}
@@ -139,6 +149,9 @@ func (daemon *Daemon) StateChanged(id string, e libcontainerd.StateInfo) error {
 		// Container is already locked in this case
 		c.Paused = false
 		daemon.setStateCounter(c)
+		if err := daemon.containersReplica.Save(c.Snapshot()); err != nil {
+			return err
+		}
 		if err := c.ToDisk(); err != nil {
 			return err
 		}

daemon/rename.go

@@ -82,6 +82,9 @@ func (daemon *Daemon) ContainerRename(oldName, newName string) error {
 		daemon.nameIndex.Release(oldName + k)
 	}
 	daemon.releaseName(oldName)
+	if err = daemon.containersReplica.Save(container.Snapshot()); err != nil {
+		return err
+	}
 	if err = container.ToDisk(); err != nil {
 		return err
 	}
@@ -99,6 +102,9 @@ func (daemon *Daemon) ContainerRename(oldName, newName string) error {
 		if err != nil {
 			container.Name = oldName
 			container.NetworkSettings.IsAnonymousEndpoint = oldIsAnonymousEndpoint
+			if e := daemon.containersReplica.Save(container.Snapshot()); err != nil {
+				logrus.Errorf("%s: Failed in replicating state on rename failure: %v", container.ID, e)
+			}
 			if e := container.ToDisk(); e != nil {
 				logrus.Errorf("%s: Failed in writing to Disk on rename failure: %v", container.ID, e)
 			}

daemon/start.go

@@ -117,8 +117,12 @@ func (daemon *Daemon) containerStart(container *container.Container, checkpoint
 			if container.ExitCode() == 0 {
 				container.SetExitCode(128)
 			}
-			container.ToDisk()
-
+			if err := daemon.containersReplica.Save(container.Snapshot()); err != nil {
+				logrus.Errorf("%s: failed replicating state on start failure: %v", container.ID, err)
+			}
+			if err := container.ToDisk(); err != nil {
+				logrus.Errorf("%s: failed writing to disk on start failure: %v", container.ID, err)
+			}
 			container.Reset(false)
 
 			daemon.Cleanup(container)

daemon/update.go

@@ -38,6 +38,7 @@ func (daemon *Daemon) update(name string, hostConfig *container.HostConfig) erro
 		if restoreConfig {
 			container.Lock()
 			container.HostConfig = &backupHostConfig
+			daemon.containersReplica.Save(container.Snapshot())
 			container.ToDisk()
 			container.Unlock()
 		}
@@ -47,10 +48,18 @@ func (daemon *Daemon) update(name string, hostConfig *container.HostConfig) erro
 		return errCannotUpdate(container.ID, fmt.Errorf("Container is marked for removal and cannot be \"update\"."))
 	}
 
+	container.Lock()
 	if err := container.UpdateContainer(hostConfig); err != nil {
 		restoreConfig = true
+		container.Unlock()
 		return errCannotUpdate(container.ID, err)
 	}
+	if err := daemon.containersReplica.Save(container.Snapshot()); err != nil {
+		restoreConfig = true
+		container.Unlock()
+		return errCannotUpdate(container.ID, err)
+	}
+	container.Unlock()
 
 	// if Restart Policy changed, we need to update container monitor
 	if hostConfig.RestartPolicy.Name != "" {