Merge pull request #35516 from mlaventure/reduce-race-on-exit

Add missing lock in ProcessEvent
Upstream-commit: 9d00efb533d3561220d169a86ca9ff965f9a6b2c
Component: engine

components/engine/daemon/monitor.go

@@ -50,12 +50,12 @@ func (daemon *Daemon) ProcessEvent(id string, e libcontainerd.EventType, ei libc
 		daemon.LogContainerEvent(c, "oom")
 	case libcontainerd.EventExit:
 		if int(ei.Pid) == c.Pid {
+			c.Lock()
 			_, _, err := daemon.containerd.DeleteTask(context.Background(), c.ID)
 			if err != nil {
 				logrus.WithError(err).Warnf("failed to delete container %s from containerd", c.ID)
 			}
 
-			c.Lock()
 			c.StreamConfig.Wait()
 			c.Reset(false)
 
@@ -72,6 +72,7 @@ func (daemon *Daemon) ProcessEvent(id string, e libcontainerd.EventType, ei libc
 				c.SetStopped(&exitStatus)
 				defer daemon.autoRemove(c)
 			}
+			defer c.Unlock() // needs to be called before autoRemove
 
 			// cancel healthcheck here, they will be automatically
 			// restarted if/when the container is started again
@@ -95,7 +96,9 @@ func (daemon *Daemon) ProcessEvent(id string, e libcontainerd.EventType, ei libc
 						}
 					}
 					if err != nil {
+						c.Lock()
 						c.SetStopped(&exitStatus)
+						c.Unlock()
 						defer daemon.autoRemove(c)
 						if err != restartmanager.ErrRestartCanceled {
 							logrus.Errorf("restartmanger wait error: %+v", err)
@@ -105,7 +108,6 @@ func (daemon *Daemon) ProcessEvent(id string, e libcontainerd.EventType, ei libc
 			}
 
 			daemon.setStateCounter(c)
-			defer c.Unlock()
 			if err := c.CheckpointTo(daemon.containersReplica); err != nil {
 				return err
 			}