You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2017/02/27 16:04:45 UTC

[jira] [Commented] (FLINK-5928) Externalized checkpoints overwritting each other

    [ https://issues.apache.org/jira/browse/FLINK-5928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15886022#comment-15886022 ] 

ASF GitHub Bot commented on FLINK-5928:
---------------------------------------

GitHub user uce opened a pull request:

    https://github.com/apache/flink/pull/3424

    [FLINK-5928] [checkpoints] Use custom metadata file for externalized checkpoints

    - Adds a checkpoint coordinator test for externalized checkpoints. This was covered only in unit tests for the involved checkpoint components like PendingCheckpoint etc. This would have caught the issue.
    - The fix is to not use a `_metadata` but a random `checkpoint_metadata-:randomSuffix` file for externalized checkpoints, because they are not unique per configured directory. Hopefully, we can get rid of this soon with @StephanEwen's refactorings.
    
    This is based on top of #3411, which is already good to merge.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/uce/flink 5928-ext_chk_metadata

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/flink/pull/3424.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #3424
    
----
commit a3d2405f690e08d3de4f641428887ab04ba2ca2d
Author: Stephan Ewen <se...@apache.org>
Date:   2017-02-17T16:51:00Z

    [FLINK-5822] [state backends] Make JobManager / Checkpoint Coordinator aware of the root state backend

commit b2f3bc41bc991e8deb22fb89822f28c75d94c8f7
Author: Stephan Ewen <se...@apache.org>
Date:   2017-02-22T21:18:50Z

    [FLINK-5897] [checkpoints] Make checkpoint externalization not depend strictly on FileSystems
    
    That is the first step towards checkpoints that can be externalized to other stores as well,
    like k/v stores and databases, if supported by the state backend.

commit 537be203dab0614383645e859bbafb6ebfeb3161
Author: Ufuk Celebi <uc...@apache.org>
Date:   2017-02-27T15:12:37Z

    [FLINK-5928] [checkpoints] Add CheckpointCoordinatorExternalizedCheckpointsTest
    
    Problem: there were only unit tests for the checkpoint instances available
    that don't test the behaviour of the checkpoint coordinator with respect
    to externalized checkpoints.

commit 88e4700cce630f8ae869abff22acfd46ab999aa0
Author: Ufuk Celebi <uc...@apache.org>
Date:   2017-02-27T15:58:14Z

    [FLINK-5928] [checkpoints] Use custom metadata file for externalized checkpoints

----


> Externalized checkpoints overwritting each other
> ------------------------------------------------
>
>                 Key: FLINK-5928
>                 URL: https://issues.apache.org/jira/browse/FLINK-5928
>             Project: Flink
>          Issue Type: Bug
>          Components: State Backends, Checkpointing
>            Reporter: Ufuk Celebi
>            Assignee: Ufuk Celebi
>            Priority: Critical
>
> I noticed that PR #3346 accidentally broke externalized checkpoints by using a fixed meta data file name. We should restore the old behaviour with creating random files and double check why no test caught this.
> This will likely superseded by upcoming changes from [~StephanEwen] to use metadata streams on the JobManager.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)