[ 
https://issues.apache.org/jira/browse/CARBONDATA-239?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15496804#comment-15496804
 ] 

ASF GitHub Bot commented on CARBONDATA-239:
-------------------------------------------

Github user gvramana commented on a diff in the pull request:

    https://github.com/apache/incubator-carbondata/pull/155#discussion_r79208835
  
    --- Diff: 
integration/spark/src/main/scala/org/apache/carbondata/spark/rdd/CarbonDataRDDFactory.scala
 ---
    @@ -590,16 +605,25 @@ object CarbonDataRDDFactory extends Logging {
                         executor, sqlContext, kettleHomePath, storeLocation
                       )
                     }
    +                catch {
    +                  case e: Exception =>
    +                    logger.error("Exception in compaction thread for table 
" + tableForCompaction
    +                      .carbonTable.getDatabaseName + "." +
    +                                 tableForCompaction.carbonTableIdentifier
    +                                   .getTableName)
    +                  // not handling the exception. only logging as this is 
not the table triggered
    +                  // by user.
    +                }
                     finally {
    -                  // delete the compaction required file
    +                  // delete the compaction required file in case of 
failure or success also.
    --- End diff --
    
    If compact file deletion fails, add failed table to skipped list.


> Failure of one compaction in queue should not affect the others.
> ----------------------------------------------------------------
>
>                 Key: CARBONDATA-239
>                 URL: https://issues.apache.org/jira/browse/CARBONDATA-239
>             Project: CarbonData
>          Issue Type: Bug
>            Reporter: ravikiran
>
> Failure of one compaction in queue should not affect the others.
> If a compaction is triggered by the user on table1 , and other requests will 
> go to queue.  and if the compaction is failed for table1 then the requests in 
> queue should continue and at the end the beeline will show the failure 
> message to the user.
> if any compaction gets failed for a table which is other than the user 
> requested table then the error in the beeline should not appear.



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

Reply via email to