[jira] [Commented] (SPARK-9498) Some statistical information missed when the driver is out of the cluster

["Liang Lee (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-9498?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14696430#comment-14696430 ] 

Liang Lee commented on SPARK-9498:
----------------------------------


Out of cluster means:
There is a cluster has three node called A,B,C.
The driver is on another node called D.
If the driver is on any of A or B or C, the info is OK.
But if the driver is on node D, some info lost as described above.
Is this clear?

> Some statistical information missed when the driver is out of the cluster
> -------------------------------------------------------------------------
>
>                 Key: SPARK-9498
>                 URL: https://issues.apache.org/jira/browse/SPARK-9498
>             Project: Spark
>          Issue Type: Improvement
>          Components: Web UI
>    Affects Versions: 1.3.1, 1.4.0
>            Reporter: Liang Lee
>
> When an application is submited and the driver is out of the Spark cluster, Some statistical information missed sometimes. 
> In stage detail inforamtion page, it will display following  info when the driver is in the spark cluster:
> Details for Stage 7
>     Total task time across all tasks: 37 min
>      Input Size / Records: 55.8 GB / 60488
>      Shuffle write: 26.6 GB / 585242962 
> But when the dreive is out of the spark cluster, it will sometimes  display above info, while sometimes not, just like this:
> Details for Stage 7
>     Total task time across all tasks: 37 min
> That is the Input Size and Shuffle data does not display.
> I have check the code and find that when the input size is zero then it will not display. 
> And the  input size is sent by each Executors and collected by Driver. 
> The problem is that the data that should be repored by Executors, is missed. 
> But I don't know why. Could anyone help to solve this problem?
>  



--
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