[
https://issues.apache.org/jira/browse/CARBONDATA-261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15507351#comment-15507351 ]
ASF GitHub Bot commented on CARBONDATA-261:
-------------------------------------------
GitHub user ravikiran23 opened a pull request:
https://github.com/apache/incubator-carbondata/pull/181 [CARBONDATA-261] clean files is updating the stale segment status.
PROBLEM:
in clean files it will read the table status first and it will delete the physical locations and update the table metadata and then take the lock for writing and after that it will write the old read list to the file.
but this is wrong as in the time between the read and write the table status would have been changed.
so before writing it should read the table status again and update the changes and then write.
SOLUTION:
before writing the status read the table status again and update the changes and then write.
You can merge this pull request into a Git repository by running:
$ git pull
https://github.com/ravikiran23/incubator-carbondata cleanfilesdefect
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-carbondata/pull/181.patchTo close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #181
----
commit 96f220f60d76b1ff63beff8516c0812c1033cd08
Author: ravikiran <
[hidden email]>
Date: 2016-09-20T18:05:17Z
in clean files it will read the table status first and it will delete the physical locations and update the table metadata and then take the lock for writing and after that it will write the old read list to the file.
but this is wrong as in the time between the read and write the table status would have been changed.
so before writing it should read the table status again and update the changes and then write.
----
> clean files is updating the stale segments to the table status.
> ---------------------------------------------------------------
>
> Key: CARBONDATA-261
> URL:
https://issues.apache.org/jira/browse/CARBONDATA-261> Project: CarbonData
> Issue Type: Bug
> Components: data-load, data-query
> Affects Versions: 0.1.0-incubating
> Reporter: ravikiran
> Assignee: ravikiran
>
> in clean files it will read the table status first and it will delete the physical locations and update the table metadata and then take the lock for writing and after that it will write the old read list to the file.
> but this is wrong as in the time between the read and write the table status would have been changed.
> so before writing it should read the table status again and update the changes and then write.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
Locked
