You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2022/05/02 05:57:55 UTC

[GitHub] [spark] sadikovi opened a new pull request, #36425: [SPARK-39084] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

sadikovi opened a new pull request, #36425:
URL: https://github.com/apache/spark/pull/36425

   <!--
   Thanks for sending a pull request!  Here are some tips for you:
     1. If this is your first time, please read our contributor guidelines: https://spark.apache.org/contributing.html
     2. Ensure you have added or run the appropriate tests for your PR: https://spark.apache.org/developer-tools.html
     3. If the PR is unfinished, add '[WIP]' in your PR title, e.g., '[WIP][SPARK-XXXX] Your PR title ...'.
     4. Be sure to keep the PR description updated to reflect all changes.
     5. Please write your PR title to summarize what this PR proposes.
     6. If possible, provide a concise example to reproduce the issue for a faster review.
     7. If you want to add a new configuration, please read the guideline first for naming configurations in
        'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
     8. If you want to add or modify an error type or message, please read the guideline first in
        'core/src/main/resources/error/README.md'.
   -->
   
   ### What changes were proposed in this pull request?
   <!--
   Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue. 
   If possible, please consider writing useful notes for better and faster reviews in your PR. See the examples below.
     1. If you refactor some codes with changing classes, showing the class hierarchy will help reviewers.
     2. If you fix some SQL features, you can provide some references of other DBMSes.
     3. If there is design documentation, please add the link.
     4. If there is a discussion in the mailing list, please add the link.
   -->
   
   This PR fixes the issue described in https://issues.apache.org/jira/browse/SPARK-39084 where calling `df.rdd.isEmpty()` on a particular dataset could result in a JVM crash and/or executor failure. 
   
   The issue was due to Python iterator not being synchronised with Java iterator so when the task is complete, the Python iterator continues to process data. We have introduced ContextAwareIterator as part of https://issues.apache.org/jira/browse/SPARK-33277 but we did not fix all of the places where this should be used.
   
   ### Why are the changes needed?
   <!--
   Please clarify why the changes are needed. For instance,
     1. If you propose a new API, clarify the use case for a new API.
     2. If you fix a bug, you can clarify why it is a bug.
   -->
   
   Fixes the JVM crash when checking isEmpty() on a dataset.
   
   ### Does this PR introduce _any_ user-facing change?
   <!--
   Note that it means *any* user-facing change including all aspects such as the documentation fix.
   If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible.
   If possible, please also clarify if this is a user-facing change compared to the released Spark versions or within the unreleased branches such as master.
   If no, write 'No'.
   -->
   
   No.
   
   ### How was this patch tested?
   <!--
   If tests were added, say they were added here. Please make sure to add some test cases that check the changes thoroughly including negative and positive cases if possible.
   If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future.
   If tests were not added, please describe why they were not added and/or why it was difficult to add.
   If benchmark tests were added, please run the benchmarks in GitHub Actions for the consistent environment, and the instructions could accord to: https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
   -->
   
   I added a test case that reproduces the issue 100%. I confirmed that the test fails without the fix and passes with the fix.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] ankurdave commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
ankurdave commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115436722

   I see, so the use-after-free that caused the crash is occurring in the main task thread? In that case this fix LGTM.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] sadikovi commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
sadikovi commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115430904

   The reason the bug existed was because the code was trying to control completion from Python which is not going to work in most of the cases.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] HyukjinKwon commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
HyukjinKwon commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115410128

   BTW I'll merge this in few days to backport to other branches as a minimal fix if there are no objection.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] HyukjinKwon closed pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
HyukjinKwon closed pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion
URL: https://github.com/apache/spark/pull/36425


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] sadikovi commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
sadikovi commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115432194

   I do have a crash file but I cannot share it unfortunately.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] sadikovi commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
sadikovi commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1114523613

   @HyukjinKwon and @ueshin Can you review the PR? Thanks.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] HyukjinKwon commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
HyukjinKwon commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1114644298

   Will leave it to @ueshin though ..


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] HyukjinKwon commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
HyukjinKwon commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115464988

   Thanks guys!!
   
   Merged to master, branch-3.3, branch-3.2, branch-3.1 and branch-3.0.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] sadikovi commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
sadikovi commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115442507

   Thanks, @ankurdave.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] ankurdave commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
ankurdave commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1116249393

   From talking to @sadikovi, it sounds like the use-after-free that caused this crash does in fact occur in the Python writer thread, not the main task thread. And since `RDD#isEmpty()` is implemented using `limit(1)`, this is a very similar situation as the one described in https://github.com/apache/spark/pull/34245. The main difference appears to be the presence of a group-by with codegen enabled.
   
   Given that, the question is why https://github.com/apache/spark/pull/34245 was not sufficient to fix this. I'm guessing [the task completion listener that frees the off-heap memory](https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala#L489) is being registered in the wrong order relative to the BasePythonRunner task completion listener.
   
   Anyway, even if that were fixed I think we would still need the fix in this PR for performance reasons. Otherwise the writer thread could read an arbitrary amount of data before checking the interrupt status.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] AmplabJenkins commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
AmplabJenkins commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1114641923

   Can one of the admins verify this patch?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] sadikovi commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
sadikovi commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115430254

   @ankurdave This PR solves just that because the synchronisation is between separate JVM threads instead which works. My test reproduces the issue and I confirm it is fixed with the patch.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] ueshin commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
ueshin commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115178840

   I remember there were some more fixes after mine.
   @ankurdave Could you also take a look at this?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org

[GitHub] [spark] ankurdave commented on pull request #36425: [SPARK-39084][PYSPARK] Fix df.rdd.isEmpty() by using TaskContext to stop iterator on task completion

Posted by GitBox <gi...@apache.org>.
ankurdave commented on PR #36425:
URL: https://github.com/apache/spark/pull/36425#issuecomment-1115248094

   As @HyukjinKwon had [noted in October 2020](https://github.com/apache/spark/pull/30177/files#r513898999), the `ContextAwareIterator` approach didn't fully solve the problem because the root issue was that there was a race between (1) the main task thread freeing an off-heap buffer, and (2) the Python WriterThread reading from that off-heap buffer. To fully solve that, it [required adding synchronization](https://github.com/apache/spark/pull/34245) between the two threads.
   
   If this issue is caused by a race between the main thread and some other thread, then it seems like the same reasoning applies. Do you have a crash file indicating which thread is causing the JVM crash?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org