[ 
https://issues.apache.org/jira/browse/HIVE-18831?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16419458#comment-16419458
 ] 

Sahil Takiar commented on HIVE-18831:
-------------------------------------

Hey [~lirui]

Thanks for the suggestion, serializing into a byte array fixed the issues I was 
seeing, the unit tests now pass!

For (2), that would make the serialization logic simpler, and we could avoid 
{{JobResultSerializer#convertToSerializableSparkException}}, but that would 
make the client side a bit messier. Specifically, the logic in {{SparkTask}} 
for parsing {[sparkJobStatus.getSparkJobException}}. If we want to enforce that 
{{getSparkJobException}} returns a {{Throwable}} then we would have to wrap the 
full-stack trace in a {{SparkException}} (which is what we do now). The issue 
is that we try to get the message from the root cause and stick in the 
{{HiveException}} message. This would cause the full stack-trace to be placed 
in the {{HiveException}} message, and the full-stack would be displayed to the 
user.

Instead of using {{SparkException}} in {{#convertToSerializableSparkException}} 
we can create something new like {{SafeSerializableExceptionWrapper}}.

> Differentiate errors that are thrown by Spark tasks
> ---------------------------------------------------
>
>                 Key: HIVE-18831
>                 URL: https://issues.apache.org/jira/browse/HIVE-18831
>             Project: Hive
>          Issue Type: Sub-task
>          Components: Spark
>            Reporter: Sahil Takiar
>            Assignee: Sahil Takiar
>            Priority: Major
>         Attachments: HIVE-18831.1.patch, HIVE-18831.2.patch, 
> HIVE-18831.3.patch, HIVE-18831.4.patch, HIVE-18831.6.patch, 
> HIVE-18831.7.patch, HIVE-18831.8.WIP.patch
>
>
> We propagate exceptions from Spark task failures to the client well, but we 
> don't differentiate between errors from HS2 / RSC vs. errors thrown by 
> individual tasks.
> Main motivation is that when the client sees a propagated Spark exception its 
> difficult to know what part of the excution threw the exception.



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

Reply via email to