[RayService][Hotfix] Hotfix for Flaky Zero Downtime Rollout Test #1837
Conversation
This file contains 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
Signed-off-by: Yicheng-Lu-llll <[email protected]>
2 tasks
kevin85421
changed the title
|
Some failures in this build appear to be related to this issue. @Yicheng-Lu-llll would you mind confirming it? Thanks! |
Signed-off-by: Yicheng-Lu-llll <[email protected]>
|
Yes, the CI failures are due to the issue mentioned here. I've included them in the sample section of my PR description. Thank you! |
kevin85421
approved these changes
Jan 17, 2024
There was a problem hiding this comment.
I haven't succeeded in reproducing the issue on my laptop. However, I encountered the issue twice when I ran the test three times in this build. After I add the hotfix to my PR, it passes the first time. Although the data points are limited, this may still somewhat validate the assumptions in this PR.
4 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Why are these changes needed?
The Zero Downtime Rollout Test is flaky for the following reason:
The current logic for the Zero Downtime Rollout operates as follows:
However, a critical issue is that the 'running' status merely indicates the readiness of the new RayCluster to serve traffic; it does not confirm that the serve service has begun redirecting traffic to the new RayCluster. As a result, requests sent after the status change to 'running' might still be responded to by the old RayCluster.
This scenario may lead to high-probability assertion errors in RayService CI tests, as responses expected from the new RayCluster may still coming from the old RayCluster. Some examples can be found here:
To mitigate this, and before refactoring the Rollout logic and redefine the of status, this PR adds delay to ensure the serve service is ready to redirect traffic to the new RayCluster.
Related issue number
Checks