runc-shim: only defer init process exits

laurazard · laurazard · commit 6d00c3ada8b0 · 2024-03-26T13:39:11.000Z
In order to make sure that we don't publish task exit events for init processes before we do for execs in that container, we added logic to `processExits` in 892dc54 to skip these and let the pending exec's `handleStarted` closure process them. However, the conditional logic in `processExits` added was faulty - we should only defer processing of exit events related to init processes, not other execs. Due to this missing condition, 892dc54 introduced a bug where, if there are many concurrent execs for the same container/init pid, exec exits are skipped and then never published, resulting in hanging clients. This commit adds the missing logic to `processExits`. Signed-off-by: Laura Brehm <laurabrehm@hey.com>
diff --git a/cmd/containerd-shim-runc-v2/task/service.go b/cmd/containerd-shim-runc-v2/task/service.go
@@ -678,7 +678,8 @@ func (s *service) processExits() {
 		// process.
 		var cps, skipped []containerProcess
 		for _, cp := range s.running[e.Pid] {
-			if s.pendingExecs[cp.Container] != 0 {
+			_, init := cp.Process.(*process.Init)
+			if init && s.pendingExecs[cp.Container] != 0 {
 				// This exit relates to a container for which we have pending execs. In
 				// order to ensure order between execs and the init process for a given
 				// container, skip processing this exit here and let the `handleStarted`
