[ 
https://issues.apache.org/jira/browse/FLINK-8802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16396910#comment-16396910
 ] 

ASF GitHub Bot commented on FLINK-8802:
---------------------------------------

GitHub user kl0u opened a pull request:

    https://github.com/apache/flink/pull/5691

    [FLINK-8802] [QS] Fix concurrent access to non-duplicated serializers.

    ## What is the purpose of the change
    
    This fixes the problem of multiple threads using the same serializer 
instance to concurrently deserialize requests.
    
    ## Brief change log
    
    Now the `KvStateRegistry` holds a `KvStateEntry` for each registered 
`InternalKvState` and in that entry we have a cache that holds a copy of the 
serializer per thread, if the serializer is stateful. If it is stateless, we do 
not need to copy the serializer.
    
    ## Verifying this change
    
    There is a test added in the `KvStateRegistryTest`, the 
`KvStateRegistryTest#testKvStateEntry()` and also I have tested it with an 
actual job (but I run it locally).
    
    ## Does this pull request potentially affect one of the following parts:
    
      - Dependencies (does it add or upgrade a dependency): no
      - The public API, i.e., is any changed class annotated with 
`@Public(Evolving)`: no
      - The serializers: no
      - The runtime per-record code paths (performance sensitive): no
      - Anything that affects deployment or recovery: JobManager (and its 
components), Checkpointing, Yarn/Mesos, ZooKeeper: no
      - The S3 file system connector: no
    
    ## Documentation
    
      - Does this pull request introduce a new feature? no
      - If yes, how is the feature documented? not applicable


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

    $ git pull https://github.com/kl0u/flink improved-qs-inv

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

    https://github.com/apache/flink/pull/5691.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 #5691
    
----
commit 595a59c30dfc778945c43351edf8dbc2887b6a87
Author: kkloudas <kkloudas@...>
Date:   2018-03-09T11:05:38Z

    [FLINK-8908] Do not create copy when MapSerializer stateless.

commit 1a8c0415cec46ed6683768b793c295ab478b2eb8
Author: kkloudas <kkloudas@...>
Date:   2018-03-12T11:12:06Z

    [FLINK-8928] [QS] Improve server binding error message.

commit 67b6d7ad2d128a14e001afc34eb14cd6c62d90c4
Author: kkloudas <kkloudas@...>
Date:   2018-03-09T21:47:35Z

    [FLINK-8802] [QS] Fix concurrent access to non-duplicated serializers.

commit b86657f630363e59971edd29464130c62bd4b3a7
Author: kkloudas <kkloudas@...>
Date:   2018-03-12T12:19:23Z

    [FLINK-8926] [QS] Shutdown client proxy after test ends.

----


> Concurrent serialization without duplicating serializers in state server.
> -------------------------------------------------------------------------
>
>                 Key: FLINK-8802
>                 URL: https://issues.apache.org/jira/browse/FLINK-8802
>             Project: Flink
>          Issue Type: Bug
>          Components: Queryable State
>    Affects Versions: 1.5.0
>            Reporter: Kostas Kloudas
>            Assignee: Kostas Kloudas
>            Priority: Blocker
>             Fix For: 1.5.0
>
>
> The `getSerializedValue()` may be called by multiple threads but serializers 
> are not duplicated, which may lead to exceptions thrown when a serializer is 
> stateful.



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

Reply via email to