beam-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From dhalperi <...@git.apache.org>
Subject [GitHub] beam pull request #2465: TestDataflowRunner: better error handling
Date Fri, 07 Apr 2017 21:56:33 GMT
GitHub user dhalperi opened a pull request:

    https://github.com/apache/beam/pull/2465

    TestDataflowRunner: better error handling

    1. There was a race in which pipelines without PAsserts might
       erroneously pass because they would be canceled, which would
       in turn cause the watermarks to reach max infinity, which
       would in turn (because there are no PAsserts) cause the main
       streaming poll loop think the pipeline succeeded.
    
       Fix this by making the error presence available to the main
       polling loop, and only canceling from there.
    
    2. The fact we were canceling from two places meant we could
       get double-cancelations that led to test failures.
    
    Fix both these issues (I hope).

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

    $ git pull https://github.com/dhalperi/beam test-dataflow-runner

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

    https://github.com/apache/beam/pull/2465.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 #2465
    
----
commit a2b509d6104a1b2b461f917773b337be388a3bfd
Author: Dan Halperin <dhalperi@google.com>
Date:   2017-04-07T21:56:09Z

    TestDataflowRunner: better error handling
    
    1. There was a race in which pipelines without PAsserts might
       erroneously pass because they would be canceled, which would
       in turn cause the watermarks to reach max infinity, which
       would in turn (because there are no PAsserts) cause the main
       streaming poll loop think the pipeline succeeded.
    
       Fix this by making the error presence available to the main
       polling loop, and only canceling from there.
    
    2. The fact we were canceling from two places meant we could
       get double-cancelations that led to test failures.
    
    Fix both these issues (I hope).

----


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message