[jira] [Commented] (CARBONDATA-244) Load and delete segment by id queries giving inconsistent results when we execute parallely

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[jira] [Commented] (CARBONDATA-244) Load and delete segment by id queries giving inconsistent results when we execute parallely

Akash R Nilugal (Jira)

    [ https://issues.apache.org/jira/browse/CARBONDATA-244?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15496915#comment-15496915 ]

ASF GitHub Bot commented on CARBONDATA-244:
-------------------------------------------

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

    https://github.com/apache/incubator-carbondata/pull/156#discussion_r79217913
 
    --- Diff: integration/spark/src/main/scala/org/apache/carbondata/spark/rdd/CarbonDataRDDFactory.scala ---
    @@ -1077,19 +1077,38 @@ object CarbonDataRDDFactory extends Logging {
           val segmentStatusManager = new SegmentStatusManager(table.getAbsoluteTableIdentifier)
           val details = segmentStatusManager
             .readLoadMetadata(loadMetadataFilePath)
    +      val carbonTableStatusLock = CarbonLockFactory
    +        .getCarbonLockObj(table.getAbsoluteTableIdentifier.getCarbonTableIdentifier,
    +          LockUsage.TABLE_STATUS_LOCK)
     
           // Delete marked loads
           val isUpdationRequired = DeleteLoadFolders
             .deleteLoadFoldersFromFileSystem(carbonLoadModel, hdfsStoreLocation,
               partitioner.partitionCount, isForceDeletion, details)
     
           if (isUpdationRequired) {
    +        try {
             // Update load metadate file after cleaning deleted nodes
    -        CarbonLoaderUtil.writeLoadMetadata(
    -          carbonLoadModel.getCarbonDataLoadSchema,
    -          carbonLoadModel.getDatabaseName,
    -          carbonLoadModel.getTableName, details.toList.asJava
    -        )
    +        if (carbonTableStatusLock.lockWithRetries()) {
    +          logger.info("Table status lock has been successfully acquired.")
    +          CarbonLoaderUtil.writeLoadMetadata(
    +            carbonLoadModel.getCarbonDataLoadSchema,
    +            carbonLoadModel.getDatabaseName,
    +            carbonLoadModel.getTableName, details.toList.asJava
    +          )
    +        }
    +        else {
    +          val errorMsg = "Clean files request is failed for " + carbonLoadModel.getDatabaseName +
    +                         "." + carbonLoadModel.getTableName +
    +                         ". Not able to acquire the table status lock."
    --- End diff --
   
    ok


> Load and delete segment by id queries giving inconsistent results when we execute parallely
> -------------------------------------------------------------------------------------------
>
>                 Key: CARBONDATA-244
>                 URL: https://issues.apache.org/jira/browse/CARBONDATA-244
>             Project: CarbonData
>          Issue Type: Bug
>            Reporter: Manohar Vanam
>            Assignee: Manohar Vanam
>
> Delete segment by id behavior is inconsistent when  we Execute load and delete segment by id queries parallely,  



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