Skip to content

testing: allow 10m for first xds fault_injection case #25779

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Mar 22, 2021
Merged

testing: allow 10m for first xds fault_injection case #25779

merged 1 commit into from
Mar 22, 2021

Conversation

dfawley
Copy link
Member

@dfawley dfawley commented Mar 22, 2021

This may be enough to fix some of the flakiness we've been seeing.

If the deletion/recreation of the backend service (to disable proxyless validation) is excessively slow, this would help. The other test that does this, circuit breaking, has a 20m timer to wait for healthy backends afterwards.

@dfawley dfawley added the release notes: no Indicates if PR should not be in release notes label Mar 22, 2021
lidizheng
lidizheng approved these changes Mar 22, 2021
View reviewed changes
Copy link
Contributor

@lidizheng lidizheng left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM to reduce impact of infrastructure flake.

Do we need to apply this extensive timeout to the very first test case for entire xDS interop run?

@dfawley
Copy link
Member Author

dfawley commented Mar 22, 2021

Do we need to apply this extensive timeout to the very first test case for entire xDS interop run?

I don't think so, unless we're seeing similar flakes with them.

@dapengzhang0
Copy link
Contributor

I don't think so, unless we're seeing similar flakes with them.

What's making fault_injection test special? I'm suspicious that if I put fault_injection test ahead of timeout test, then timeout test might be flaky.

@dfawley
Copy link
Member Author

dfawley commented Mar 22, 2021

What's making fault_injection test special? I'm suspicious that if I put fault_injection test ahead of timeout test, then timeout test might be flaky.

I think it might be the disabling of proxyless validation, which leads to the whole backend service being deleted and recreated. The timeout test doesn't do that. circuit_breaking does, but it has a couple wait_for_healthy_backends afterwords, which have 10m timeouts.

@dfawley dfawley merged commit e5829de into grpc:master Mar 22, 2021
@dfawley dfawley deleted the longer_timeout_fi branch March 22, 2021 23:02
@dfawley dfawley mentioned this pull request Mar 22, 2021
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
3 more reviewers

@dapengzhang0 dapengzhang0 dapengzhang0 approved these changes

@menghanl menghanl menghanl approved these changes

@lidizheng lidizheng lidizheng approved these changes

Reviewers whose approvals may not affect merge requirements
Assignees

@dapengzhang0 dapengzhang0

@lidizheng lidizheng

Labels
release notes: no Indicates if PR should not be in release notes
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@dfawley @dapengzhang0 @menghanl @lidizheng