[SPARK-49983][CORE][TESTS] Fix BarrierTaskContextSuite.successively sync with allGather and barrier test case to be robust
#48487
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
…sync with allGather and barrier` test case to be robust
xinrong-meng
approved these changes
Oct 16, 2024
Member
|
LGTM, the tests in Build (pull_request_target) all passed. Thank you! |
Member
Author
|
Thank you so much, @xinrong-meng . |
dongjoon-hyun
added a commit
that referenced
this pull request
Oct 16, 2024
…sync with allGather and barrier` test case to be robust ### What changes were proposed in this pull request? This PR aims to fix `BarrierTaskContextSuite.successively sync with allGather and barrier` test case to be robust. ### Why are the changes needed? The test case asserts the duration of partitions. However, this is flaky because we don't know when a partition is triggered before `barrier` sync. https://github.com/apache/spark/blob/0e75d19a736aa18fe77414991ebb7e3577a43af8/core/src/test/scala/org/apache/spark/scheduler/BarrierTaskContextSuite.scala#L116-L118 Although we added `TestUtils.waitUntilExecutorsUp` at Apache Spark 3.0.0 like the following, - #28658 let's say a partition starts slowly than `38ms` and all partitions sleep `1s` exactly. Then, the test case fails like the following. - https://github.com/apache/spark/actions/runs/11298639789/job/31428018075 ``` BarrierTaskContextSuite: ... - successively sync with allGather and barrier *** FAILED *** 1038 was not less than or equal to 1000 (BarrierTaskContextSuite.scala:118) ``` According to the failure history here (SPARK-49983) and SPARK-31730, the slowness seems to be less than `200ms` when it happens. So, this PR aims to reduce the flakiness by capping the sleep up to 500ms while keeping the `1s` validation. There is no test coverage change because this test case focuses on the `successively sync with allGather and battier`. ### Does this PR introduce _any_ user-facing change? No, this is a test-only test case. ### How was this patch tested? Pass the CIs. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #48487 from dongjoon-hyun/SPARK-49983. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit bcfe62b) Signed-off-by: Dongjoon Hyun <[email protected]>
dongjoon-hyun
added a commit
that referenced
this pull request
Oct 16, 2024
…sync with allGather and barrier` test case to be robust ### What changes were proposed in this pull request? This PR aims to fix `BarrierTaskContextSuite.successively sync with allGather and barrier` test case to be robust. ### Why are the changes needed? The test case asserts the duration of partitions. However, this is flaky because we don't know when a partition is triggered before `barrier` sync. https://github.com/apache/spark/blob/0e75d19a736aa18fe77414991ebb7e3577a43af8/core/src/test/scala/org/apache/spark/scheduler/BarrierTaskContextSuite.scala#L116-L118 Although we added `TestUtils.waitUntilExecutorsUp` at Apache Spark 3.0.0 like the following, - #28658 let's say a partition starts slowly than `38ms` and all partitions sleep `1s` exactly. Then, the test case fails like the following. - https://github.com/apache/spark/actions/runs/11298639789/job/31428018075 ``` BarrierTaskContextSuite: ... - successively sync with allGather and barrier *** FAILED *** 1038 was not less than or equal to 1000 (BarrierTaskContextSuite.scala:118) ``` According to the failure history here (SPARK-49983) and SPARK-31730, the slowness seems to be less than `200ms` when it happens. So, this PR aims to reduce the flakiness by capping the sleep up to 500ms while keeping the `1s` validation. There is no test coverage change because this test case focuses on the `successively sync with allGather and battier`. ### Does this PR introduce _any_ user-facing change? No, this is a test-only test case. ### How was this patch tested? Pass the CIs. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #48487 from dongjoon-hyun/SPARK-49983. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit bcfe62b) Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit d37a8b9) Signed-off-by: Dongjoon Hyun <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
What changes were proposed in this pull request?
This PR aims to fix
BarrierTaskContextSuite.successively sync with allGather and barriertest case to be robust.Why are the changes needed?
The test case asserts the duration of partitions. However, this is flaky because we don't know when a partition is triggered before
barriersync.spark/core/src/test/scala/org/apache/spark/scheduler/BarrierTaskContextSuite.scala
Lines 116 to 118 in 0e75d19
Although we added
TestUtils.waitUntilExecutorsUpat Apache Spark 3.0.0 like the following,let's say a partition starts slowly than
38msand all partitions sleep1sexactly. Then, the test case fails like the following.According to the failure history here (SPARK-49983) and SPARK-31730, the slowness seems to be less than
200mswhen it happens. So, this PR aims to reduce the flakiness by capping the sleep up to 500ms while keeping the1svalidation. There is no test coverage change because this test case focuses on thesuccessively sync with allGather and battier.Does this PR introduce any user-facing change?
No, this is a test-only test case.
How was this patch tested?
Pass the CIs.
Was this patch authored or co-authored using generative AI tooling?
No.