Merge pull request #16111 from vrothberg/healthcheck-fixes

healthcheck: fix --on-failure=stop

Author: openshift-merge-robot

libpod/container_internal.go

@@ -1285,12 +1285,6 @@ func (c *Container) stop(timeout uint) error {
 		c.lock.Unlock()
 	}
 
-	if c.config.HealthCheckConfig != nil {
-		if err := c.removeTransientFiles(context.Background()); err != nil {
-			logrus.Error(err.Error())
-		}
-	}
-
 	stopErr := c.ociRuntime.StopContainer(c, timeout, all)
 
 	if !c.batched {
@@ -1415,6 +1409,11 @@ func (c *Container) restartWithTimeout(ctx context.Context, timeout uint) (retEr
 		if err := c.stop(timeout); err != nil {
 			return err
 		}
+		if c.config.HealthCheckConfig != nil {
+			if err := c.removeTransientFiles(context.Background()); err != nil {
+				logrus.Error(err.Error())
+			}
+		}
 		// Old versions of conmon have a bug where they create the exit file before
 		// closing open file descriptors causing a race condition when restarting
 		// containers with open ports since we cannot bind the ports as they're not

test/system/220-healthcheck.bats

@@ -123,6 +123,8 @@ Log[-1].Output   | \"Uh-oh on stdout!\\\nUh-oh on stderr!\"
 	    # kill and stop yield the container into a non-running state
             is "$output" ".* $policy" "container was stopped/killed"
             assert "$output" != "running $policy"
+            # also make sure that it's not stuck in the stopping state
+            assert "$output" != "stopping $policy"
         fi
 
         run_podman rm -f -t0 $ctr