Ensure graceful shutdown in GenericContainer.stop() (#10365) #11350
+46
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Issue #10365
Motivation
This PR addresses Issue #10365 where the trap signal in a container script is ignored when GenericContainer#stop() is called.
The root cause is that GenericContainer#stop() currently delegates the stop process directly to ResourceReaper#stopAndRemoveContainer(), which issues a killContainerCmd (SIGKILL). This prevents the container from performing any graceful shutdown tasks (e.g., closing connections, flushing logs, executing hooks).
This PR modifies GenericContainer#stop() to attempt a dockerClient.stopContainerCmd() (SIGTERM) before delegating removal to the ResourceReaper.
Changes
Modified GenericContainer#stop() to issue stopContainerCmd before calling ResourceReaper.
Added testPostHookExecutionOnStop in GenericContainerTest to verify that trap hooks are correctly executed upon stopping.
Discussion / Concerns (Reason for Draft)
I am submitting this as a Draft because changing the default behavior to stopContainerCmd implies that stop() will now wait for the container to exit.
Pros: Correctness. Containers can handle lifecycle hooks properly.
Cons: Potential performance regression in tests where containers ignore SIGTERM, causing a delay of up to 10s (default timeout) per test.
I have verified locally that well-behaved containers (which handle SIGTERM) stop almost instantly, so the performance impact should be minimal for correct implementations. However, I would like to hear the maintainers' opinion on whether this should be the default behavior or if we should introduce an opt-in configuration (e.g., .withShutdownTimeout()).