shimv2: kill a container return directly once the container termianted

According to CRI specs, kubelet will call StopPodSandbox() at least once before calling RemovePodSandbox, and this call is idempotent, and must not return an error if all relevant resources have already been reclaimed. And in that call it will send a SIGKILL signal first to try to stop the container, thus once the container has terminated, here should ignore this signal and return directly. Fixes:#1672 Signed-off-by: lifupan <[email protected]> (cherry picked from commit 0d535f5) Signed-off-by: Ganesh Maharaj Mahalingam <[email protected]>
kata-containers · Jun 5, 2019 · a38b402 · a38b402
1 parent 377f332
commit a38b402
Showing 1 changed file with 14 additions and 0 deletions.
diff --git a/containerd-shim-v2/service.go b/containerd-shim-v2/service.go
@@ -678,6 +678,20 @@ func (s *service) Kill(ctx context.Context, r *taskAPI.KillRequest) (_ *ptypes.E
 		return nil, err
 	}
 
+	// According to CRI specs, kubelet will call StopPodSandbox()
+	// at least once before calling RemovePodSandbox, and this call
+	// is idempotent, and must not return an error if all relevant
+	// resources have already been reclaimed. And in that call it will
+	// send a SIGKILL signal first to try to stop the container, thus
+	// once the container has terminated, here should ignore this signal
+	// and return directly.
+	if signum == syscall.SIGKILL || signum == syscall.SIGTERM {
+		if c.status == task.StatusStopped {
+			logrus.WithField("sandbox", s.sandbox.ID()).WithField("Container", c.id).Debug("Container has already been stopped")
+			return empty, nil
+		}
+	}
+
 	processID := c.id
 	if r.ExecID != "" {
 		execs, err := c.getExec(r.ExecID)