Github user JoshRosen commented on the issue:

    https://github.com/apache/spark/pull/20310
  
    Are you sure that we want to blanket revert this entire patch? Is there a 
more surgical short-term fix we can make in `dev/sparktestsupport/modules.py` 
to just always unconditionally enable the tag for now?
    
    Also, is this the first time recently that we've failed the YARN 
integration tests? How much time do they add?
    
    The trade off here seems to be between slightly slower after-the-fact 
detection of a test failure / build break due to YARN vs. faster tests for the 
majority of PRs that don't touch YARN code. I think we've had one or two such 
breaks in the 2+ years that we've been using these test tags, so I'd also be 
fine with postponing this change if you agree that it's unlikely that we're 
going to have many such failures here.
    
    If the motivation is that it's hard to test the fix for such build breaks 
(because the failing test wouldn't be exercised in the PR builder) then I think 
we might already have a solution via special tags placed into the PR title (I 
think `test-yarn` or something similar; see `run-tests-jenkins.py`).


---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org

Reply via email to