contrib/checkpoint: increase timeouts to 30s#13436
Conversation
Under slow network conditions (e.g., simulated 2 Mbps ingress bandwidth limits in CI), restoring a container from a checkpoint can fail if it requires pulling the base image again. During restore, crictl create is called. In test_from_oci, it used a 10-second timeout (crictl -t 10s create). Over a 2 Mbps connection, pulling the 2.8MB Alpine layer takes ~12-13 seconds, causing the operation to exceed the timeout and fail with context canceled. Increase the timeouts in both test_from_archive and test_from_oci to 30 seconds. This ensures the base image pull can complete successfully even in constrained CI environments. Tested: Ran the CRI integration tests on a GCE VM under a simulated 2 Mbps ingress bandwidth limit using traffic control (tc). Verified that both test_from_archive and test_from_oci passed successfully and container restore completed without timeout. Assisted-by: Antigravity Signed-off-by: Samuel Karp <[email protected]>
There was a problem hiding this comment.
Pull request overview
Adjusts CRI checkpoint/restore integration helper timeouts to better tolerate slow network pulls during restore, improving reliability of the contrib checkpoint restore tests in constrained CI environments.
Changes:
- Increase
crictl createtimeout intest_from_archivefrom 20s to 30s. - Increase
crictl createtimeout intest_from_ocifrom 10s to 30s.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
github-project-automation
Bot
moved this from Needs Triage
to Review In Progress
in Pull Request Review
github-project-automation
Bot
moved this from Review In Progress
to Done
in Pull Request Review
|
/cherry-pick release/2.3 |
|
/cherry-pick release/2.2 |
|
@samuelkarp: new pull request created: #13459 DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
@samuelkarp: new pull request created: #13460 DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
samuelkarp
added
cherry-pick/2.2.x
Under slow network conditions (e.g., simulated 2 Mbps ingress bandwidth limits in CI), restoring a container from a checkpoint can fail if it requires pulling the base image again.
During restore, crictl create is called. In test_from_oci, it used a 10-second timeout (crictl -t 10s create). Over a 2 Mbps connection, pulling the 2.8MB Alpine layer takes ~12-13 seconds, causing the operation to exceed the timeout and fail with context canceled.
Increase the timeouts in both test_from_archive and test_from_oci to 30 seconds. This ensures the base image pull can complete successfully even in constrained CI environments.
Tested:
Ran the CRI integration tests on a GCE VM under a simulated 2 Mbps ingress bandwidth limit using traffic control (tc). Verified that both test_from_archive and test_from_oci passed successfully and container restore completed without timeout.
Assisted-by: Antigravity