[jira] [Updated] (SPARK-23811) FetchFailed comes before Success of same task will cause child stage never succeed

2018-03-30 Thread Li Yuanjian (JIRA)

 [ 
https://issues.apache.org/jira/browse/SPARK-23811?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Li Yuanjian updated SPARK-23811:

Description: 
This is a bug caused by abnormal scenario describe below:
 # ShuffleMapTask 1.0 running, this task will fetch data from ExecutorA
 # ExecutorA Lost, trigger `mapOutputTracker.removeOutputsOnExecutor(execId)` , 
shuffleStatus changed.
 # Speculative ShuffleMapTask 1.1 start, got a FetchFailed immediately.
 # ShuffleMapTask 1.0 finally succeed, but because of 1.1's FetchFailed, stage 
still mark as failed stage.
 # ShuffleMapTask 1 is the last task of its stage, so this stage will never 
succeed because of there's no missing task DagScheduler can get.

  was:
This is a bug caused by abnormal scenario describe below:
 # ShuffleMapTask 1.0 running, this task will fetch data from ExecutorA
 # ExecutorA Lost, trigger `mapOutputTracker.removeOutputsOnExecutor(execId)` , 
shuffleStatus changed.
 # Speculative ShuffleMapTask 1.1 start, got a FetchFailed immediately.
 # ShuffleMapTask 1 is the last task of its stage, so this stage will never 
succeed because of there's no missing task DagScheduler can get.


> FetchFailed comes before Success of same task will cause child stage never 
> succeed
> --
>
> Key: SPARK-23811
> URL: https://issues.apache.org/jira/browse/SPARK-23811
> Project: Spark
>  Issue Type: Bug
>  Components: Spark Core
>Affects Versions: 2.2.0, 2.3.0
>Reporter: Li Yuanjian
>Priority: Major
> Attachments: 1.png, 2.png
>
>
> This is a bug caused by abnormal scenario describe below:
>  # ShuffleMapTask 1.0 running, this task will fetch data from ExecutorA
>  # ExecutorA Lost, trigger `mapOutputTracker.removeOutputsOnExecutor(execId)` 
> , shuffleStatus changed.
>  # Speculative ShuffleMapTask 1.1 start, got a FetchFailed immediately.
>  # ShuffleMapTask 1.0 finally succeed, but because of 1.1's FetchFailed, 
> stage still mark as failed stage.
>  # ShuffleMapTask 1 is the last task of its stage, so this stage will never 
> succeed because of there's no missing task DagScheduler can get.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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



[jira] [Updated] (SPARK-23811) FetchFailed comes before Success of same task will cause child stage never succeed

2018-03-29 Thread Li Yuanjian (JIRA)

 [ 
https://issues.apache.org/jira/browse/SPARK-23811?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Li Yuanjian updated SPARK-23811:

Summary: FetchFailed comes before Success of same task will cause child 
stage never succeed  (was: Same tasks' FetchFailed event comes before Success 
will cause child stage never succeed)

> FetchFailed comes before Success of same task will cause child stage never 
> succeed
> --
>
> Key: SPARK-23811
> URL: https://issues.apache.org/jira/browse/SPARK-23811
> Project: Spark
>  Issue Type: Bug
>  Components: Spark Core
>Affects Versions: 2.2.0, 2.3.0
>Reporter: Li Yuanjian
>Priority: Major
> Attachments: 1.png, 2.png
>
>
> This is a bug caused by abnormal scenario describe below:
>  # ShuffleMapTask 1.0 running, this task will fetch data from ExecutorA
>  # ExecutorA Lost, trigger `mapOutputTracker.removeOutputsOnExecutor(execId)` 
> , shuffleStatus changed.
>  # Speculative ShuffleMapTask 1.1 start, got a FetchFailed immediately.
>  # ShuffleMapTask 1 is the last task of its stage, so this stage will never 
> succeed because of there's no missing task DagScheduler can get.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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