GitHub user dongjoon-hyun opened a pull request:

    https://github.com/apache/spark/pull/22150

     [SPARK-25144][SQL][TEST][BRANCH-2.3] Free aggregate map when task ends

    ## What changes were proposed in this pull request?
    
    [SPARK-25144](https://issues.apache.org/jira/browse/SPARK-25144) reports 
memory leaks on Apache Spark 2.1.3, 2.2.2, 2.3.2-RC5.
    
    ```scala
    scala> case class Foo(bar: Option[String])
    scala> val ds = List(Foo(Some("bar"))).toDS
    scala> val result = ds.flatMap(_.bar).distinct
    scala> result.rdd.isEmpty
    18/08/19 23:01:54 WARN Executor: Managed memory leak detected; size = 
8650752 bytes, TID = 125
    res0: Boolean = false
    ```
    
    This is a backport of @cloud-fan 's 
https://github.com/apache/spark/pull/21738 which is a single commit among 3 
commits of SPARK-21743. In addition, I added a test case to prevent regressions 
in branch-2.3 and branch-2.2. Although SPARK-21743 is reverted due to 
regression, this subpatch can go to branch-2.3 and branch-2.2.
    
    ## How was this patch tested?
    
    Pass the jenkins with a newly added test case.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/dongjoon-hyun/spark SPARK-25144

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/spark/pull/22150.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #22150
    
----
commit 4355bf7f49b94dcfcfbe9df6be4c5c66cd0144e4
Author: Wenchen Fan <wenchen@...>
Date:   2018-07-10T15:07:10Z

    [SPARK-25144][SQL] Free aggregate map when task ends
    
    This is a backport of https://github.com/apache/spark/pull/21738.
    
    Author: Wenchen Fan <[email protected]>
    
    Closes #21738 from cloud-fan/limit.

commit 5f7f4466fb6192bc1b45d76e7c3a514d75e482ab
Author: Dongjoon Hyun <dongjoon@...>
Date:   2018-08-20T05:56:22Z

    [SPARK-25144][SQL][TEST][BRANCH-2.3] Add a test case.

----


---

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to