GitHub user pnowojski opened a pull request:

    https://github.com/apache/flink/pull/4955

    [FLINK-7978][kafka] Ensure that transactional ids will never clash

    ## What is the purpose of the change
    
    Previously transactional ids to use and to abort could clash between 
subtasks. This could lead to a race condition between initialization and 
writting the data, where one subtask is still initializing/aborting some 
transactional id while different subtask is already trying to write the data 
using the same transactional id.
    
    ## Brief change log
    
    First commit extracts `TransactionalIdsGenerator` logic and is a pure 
refactor, without any functional change. Second one is the actual fix. I would 
like to merge those two commits as they are, without squashing, so that a bug 
fix is actually easier to understand/read in the commit history.
    
    ## Verifying this change
    
    This fixes a bug and adds test coverage (`TransactionalIdsGeneratorTest`) 
for future regressions.
    
    ## Does this pull request potentially affect one of the following parts:
    
      - Dependencies (does it add or upgrade a dependency): (yes / **no**)
      - The public API, i.e., is any changed class annotated with 
`@Public(Evolving)`: (yes / **no**)
      - The serializers: (yes / **no** / don't know)
      - The runtime per-record code paths (performance sensitive): (yes / 
**no** / don't know)
      - Anything that affects deployment or recovery: JobManager (and its 
components), Checkpointing, Yarn/Mesos, ZooKeeper: (yes / **no** / don't know)
      - The S3 file system connector: (yes / **no** / don't know)
    
    ## Documentation
    
      - Does this pull request introduce a new feature? (yes / **no**)
      - If yes, how is the feature documented? (**not applicable** / docs / 
JavaDocs / not documented)


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

    $ git pull https://github.com/pnowojski/flink f7978

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

    https://github.com/apache/flink/pull/4955.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 #4955
    
----
commit 54c1729fed93e51f416d365c65cbd64f67c27321
Author: Piotr Nowojski <piotr.nowoj...@gmail.com>
Date:   2017-11-06T13:03:16Z

    [hotfix][kafka] Extract TransactionalIdsGenerator class from 
FlinkKafkaProducer011
    
    This is pure refactor without any functional changes.

commit 96bb26e5d9289fb8393f0807936792d89138a744
Author: Piotr Nowojski <piotr.nowoj...@gmail.com>
Date:   2017-11-06T13:14:01Z

    [FLINK-7978][kafka] Ensure that transactional ids will never clash
    
    Previously transactional ids to use and to abort could clash between
    subtasks. This could lead to a race condition between initialization
    and writting the data, where one subtask is still initializing/aborting
    some transactional id while different subtask is already trying to write
    the data using the same transactional id.

----


---

Reply via email to