[jira] [Updated] (SPARK-14911) Fix a potential data race in TaskMemoryManager

["Davies Liu (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SPARK-14911?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Davies Liu updated SPARK-14911:
-------------------------------
    Assignee: Liwei Lin

> Fix a potential data race in TaskMemoryManager
> ----------------------------------------------
>
>                 Key: SPARK-14911
>                 URL: https://issues.apache.org/jira/browse/SPARK-14911
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core
>    Affects Versions: 2.0.0
>            Reporter: Liwei Lin
>            Assignee: Liwei Lin
>            Priority: Minor
>             Fix For: 2.0.0
>
>
> SPARK-13210 introduced an `acquiredButNotUsed` field, but it might not be correctly synchronized:
> - the write `acquiredButNotUsed += acquired` is guarded by `this` lock (see [here|https://github.com/apache/spark/blame/master/core/src/main/java/org/apache/spark/memory/TaskMemoryManager.java#L271]);
> - the read `memoryManager.releaseExecutionMemory(acquiredButNotUsed, taskAttemptId, tungstenMemoryMode)` (see [here|https://github.com/apache/spark/blame/master/core/src/main/java/org/apache/spark/memory/TaskMemoryManager.java#L400]) might not be correctly synchronized, and might not see `acquiredButNotUsed`'s new written value.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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