[SPARK-28247][SS][TEST]Fix flaky test "query without test harness" on ContinuousSuite #32316
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
Member
Author
|
cc @jose-torres |
|
Kubernetes integration test unable to build dist. exiting with code: 1 |
|
Test build #137872 has finished for PR 32316 at commit
|
Member
|
It seems to fail Scala 2.13 build.
|
|
Kubernetes integration test starting |
|
Kubernetes integration test status failure |
|
Test build #137896 has finished for PR 32316 at commit
|
viirya
approved these changes
Apr 25, 2021
HyukjinKwon
approved these changes
Apr 25, 2021
Member
|
cc @HeartSaVioR FYI |
HeartSaVioR
approved these changes
Apr 26, 2021
Contributor
HeartSaVioR
left a comment
HeartSaVioR
left a comment
There was a problem hiding this comment.
LGTM, thanks for the fix!
Contributor
|
Would we like to wait for @jose-torres to do the final review (and probably sign-off), or OK to go merging? |
dongjoon-hyun
approved these changes
Apr 26, 2021
Contributor
|
OK thanks everyone for reviewing. I'm going to merge this. |
HeartSaVioR
pushed a commit
that referenced
this pull request
Apr 26, 2021
…n ContinuousSuite ### What changes were proposed in this pull request? This is another attempt to fix the flaky test "query without test harness" on ContinuousSuite. `query without test harness` is flaky because it starts a continuous query with two partitions but assumes they will run at the same speed. In this test, 0 and 2 will be written to partition 0, 1 and 3 will be written to partition 1. It assumes when we see 3, 2 should be written to the memory sink. But this is not guaranteed. We can add `if (currentValue == 2) Thread.sleep(5000)` at this line https://github.com/apache/spark/blob/b2a2b5d8206b7c09b180b8b6363f73c6c3fdb1d8/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/continuous/ContinuousRateStreamSource.scala#L135 to reproduce the failure: `Result set Set([0], [1], [3]) are not a superset of Set(0, 1, 2, 3)!` The fix is changing `waitForRateSourceCommittedValue` to wait until all partitions reach the desired values before stopping the query. ### Why are the changes needed? Fix a flaky test. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Existing tests. Manually verify the reproduction I mentioned above doesn't fail after this change. Closes #32316 from zsxwing/SPARK-28247-fix. Authored-by: Shixiong Zhu <[email protected]> Signed-off-by: Jungtaek Lim <[email protected]> (cherry picked from commit 0df3b50) Signed-off-by: Jungtaek Lim <[email protected]>
HeartSaVioR
pushed a commit
that referenced
this pull request
Apr 26, 2021
…n ContinuousSuite ### What changes were proposed in this pull request? This is another attempt to fix the flaky test "query without test harness" on ContinuousSuite. `query without test harness` is flaky because it starts a continuous query with two partitions but assumes they will run at the same speed. In this test, 0 and 2 will be written to partition 0, 1 and 3 will be written to partition 1. It assumes when we see 3, 2 should be written to the memory sink. But this is not guaranteed. We can add `if (currentValue == 2) Thread.sleep(5000)` at this line https://github.com/apache/spark/blob/b2a2b5d8206b7c09b180b8b6363f73c6c3fdb1d8/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/continuous/ContinuousRateStreamSource.scala#L135 to reproduce the failure: `Result set Set([0], [1], [3]) are not a superset of Set(0, 1, 2, 3)!` The fix is changing `waitForRateSourceCommittedValue` to wait until all partitions reach the desired values before stopping the query. ### Why are the changes needed? Fix a flaky test. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Existing tests. Manually verify the reproduction I mentioned above doesn't fail after this change. Closes #32316 from zsxwing/SPARK-28247-fix. Authored-by: Shixiong Zhu <[email protected]> Signed-off-by: Jungtaek Lim <[email protected]> (cherry picked from commit 0df3b50) Signed-off-by: Jungtaek Lim <[email protected]>
Contributor
|
Thanks @zsxwing for the fix! I merged this in master/3.1/3.0. I skipped 2.4 as it's unlikely that we'll want to maintain 2.4 version line further. |
jdcasale
pushed a commit
to palantir/spark
that referenced
this pull request
Jun 22, 2021
…n ContinuousSuite ### What changes were proposed in this pull request? This is another attempt to fix the flaky test "query without test harness" on ContinuousSuite. `query without test harness` is flaky because it starts a continuous query with two partitions but assumes they will run at the same speed. In this test, 0 and 2 will be written to partition 0, 1 and 3 will be written to partition 1. It assumes when we see 3, 2 should be written to the memory sink. But this is not guaranteed. We can add `if (currentValue == 2) Thread.sleep(5000)` at this line https://github.com/apache/spark/blob/b2a2b5d8206b7c09b180b8b6363f73c6c3fdb1d8/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/continuous/ContinuousRateStreamSource.scala#L135 to reproduce the failure: `Result set Set([0], [1], [3]) are not a superset of Set(0, 1, 2, 3)!` The fix is changing `waitForRateSourceCommittedValue` to wait until all partitions reach the desired values before stopping the query. ### Why are the changes needed? Fix a flaky test. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Existing tests. Manually verify the reproduction I mentioned above doesn't fail after this change. Closes apache#32316 from zsxwing/SPARK-28247-fix. Authored-by: Shixiong Zhu <[email protected]> Signed-off-by: Jungtaek Lim <[email protected]>
flyrain
pushed a commit
to flyrain/spark
that referenced
this pull request
Sep 21, 2021
…n ContinuousSuite ### What changes were proposed in this pull request? This is another attempt to fix the flaky test "query without test harness" on ContinuousSuite. `query without test harness` is flaky because it starts a continuous query with two partitions but assumes they will run at the same speed. In this test, 0 and 2 will be written to partition 0, 1 and 3 will be written to partition 1. It assumes when we see 3, 2 should be written to the memory sink. But this is not guaranteed. We can add `if (currentValue == 2) Thread.sleep(5000)` at this line https://github.com/apache/spark/blob/b2a2b5d8206b7c09b180b8b6363f73c6c3fdb1d8/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/continuous/ContinuousRateStreamSource.scala#L135 to reproduce the failure: `Result set Set([0], [1], [3]) are not a superset of Set(0, 1, 2, 3)!` The fix is changing `waitForRateSourceCommittedValue` to wait until all partitions reach the desired values before stopping the query. ### Why are the changes needed? Fix a flaky test. ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Existing tests. Manually verify the reproduction I mentioned above doesn't fail after this change. Closes apache#32316 from zsxwing/SPARK-28247-fix. Authored-by: Shixiong Zhu <[email protected]> Signed-off-by: Jungtaek Lim <[email protected]> (cherry picked from commit 0df3b50) Signed-off-by: Jungtaek Lim <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 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 is another attempt to fix the flaky test "query without test harness" on ContinuousSuite.
query without test harnessis flaky because it starts a continuous query with two partitions but assumes they will run at the same speed.In this test, 0 and 2 will be written to partition 0, 1 and 3 will be written to partition 1. It assumes when we see 3, 2 should be written to the memory sink. But this is not guaranteed. We can add
if (currentValue == 2) Thread.sleep(5000)at this linespark/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/continuous/ContinuousRateStreamSource.scala
Line 135 in b2a2b5d
Result set Set([0], [1], [3]) are not a superset of Set(0, 1, 2, 3)!The fix is changing
waitForRateSourceCommittedValueto wait until all partitions reach the desired values before stopping the query.Why are the changes needed?
Fix a flaky test.
Does this PR introduce any user-facing change?
No
How was this patch tested?
Existing tests. Manually verify the reproduction I mentioned above doesn't fail after this change.