roachtest: improve error reporting for tests that time out #114565
Labels
A-testing
Testing tools and infrastructure
C-enhancement
Solution expected to add code/behavior + preserve backward-compat (pg compat issues are exception)
E-quick-win
Likely to be a quick win for someone experienced.
T-testeng
TestEng Team
Comments
renatolabs
added
C-enhancement
![blathers[crl]](https://avatars.githubusercontent.com/in/59700?s=40&u=373d9101a12edc3847d2fb6ae78e3adcfb4ca41e&v=4){kind=avatar}
|
cc @cockroachdb/test-eng |
renatolabs
added
the
E-quick-win
craig bot
pushed a commit
that referenced
this issue
Dec 1, 2023
115286: roachtest: no longer show other errors if a test times out r=renatolabs a=DarrylWong When a test times out, it often leads to other failures (i.e. context cancelled). However, these are often just noise and distract from the root cause of the failures, the timeout. This change makes it so when a timeout failure is added, it suppresses future errors from being shown in github posts and the test failure message. They are still logged in their respective failure.log files. Epic: none Release note: none Fixes: #114565 Co-authored-by: DarrylWong <darryl@cockroachlabs.com>
blathers-crl bot
pushed a commit
that referenced
this issue
Dec 1, 2023
When a test times out, it often leads to other failures (i.e. context cancelled). However, these are often just noise and distract from the root cause of the failures, the timeout. This change makes it so when a timeout failure is added, it suppresses future errors from being shown in github posts and the test failure message. They are still logged in their respective failure.log files. Epic: none Release note: none Fixes: #114565
blathers-crl bot
pushed a commit
that referenced
this issue
Dec 1, 2023
When a test times out, it often leads to other failures (i.e. context cancelled). However, these are often just noise and distract from the root cause of the failures, the timeout. This change makes it so when a timeout failure is added, it suppresses future errors from being shown in github posts and the test failure message. They are still logged in their respective failure.log files. Epic: none Release note: none Fixes: #114565
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When a test times out, it's often that case that something else also fails: the context is canceled, which causes the cockroach processes to die, and tests often see errors as a consequence of that. This makes the current error reporting logic confusing: multiple errors are displayed in the GitHub issue, often taking attention away from the fact that the test timed out.
A recent example is #114501.
Error reporting could use a lot of improvements, but an immediate change we could make here is to not display any other errors if the test timed out, as that is always going to be the main error we should pay attention to.
Jira issue: CRDB-33567
The text was updated successfully, but these errors were encountered: