GitHub user zentol opened a pull request:

    https://github.com/apache/flink/pull/5657

    [FLINK-8887][tests] Add single retry in MiniClusterClient

    ## What is the purpose of the change
    
    This PR presents a test workaround for race-conditions in FLIP-6 (most 
notably FLINK-8887). Basically, every `MiniClusterClient` call is retried 
*once* after 500ms in case of certain exceptions.
    
    **This is only a band-aid until a proper fix is in place** so we can 
finally continue merging more test ports.
    
    ## Brief change log
    
    * add `guardWithSingleRetry` convenience method
    * add `ScheduledExecutor` to `MiniClusterClient`
    * guard all calls to the `MiniCluster`
    
    ## Verifying this change
    
    The change can be verified by cherry-picking [this 
branch](https://github.com/zentol/flink/tree/8797) and running the 
`AbstractOperatorRestoreTestBase`. Before this change there was always 1-2 
tests failing, whereas now none should fail.
    
    /cc @aljoscha @GJL 

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

    $ git pull https://github.com/zentol/flink 8887_bandaid

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

    https://github.com/apache/flink/pull/5657.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 #5657
    
----
commit f685fad6731b7c1774b247985a446260ea285663
Author: zentol <chesnay@...>
Date:   2018-03-07T14:02:05Z

    [FLINK-8887][tests] Add single retry in MiniClusterClient

----


---

Reply via email to