zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ZOOKEEPER-2786) Flaky test: org.apache.zookeeper.test.ClientTest.testNonExistingOpCode
Date Thu, 03 Aug 2017 20:12:01 GMT

    [ https://issues.apache.org/jira/browse/ZOOKEEPER-2786?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16113399#comment-16113399
] 

ASF GitHub Bot commented on ZOOKEEPER-2786:
-------------------------------------------

GitHub user afine opened a pull request:

    https://github.com/apache/zookeeper/pull/327

    ZOOKEEPER-2786 Flaky test: org.apache.zookeeper.test.ClientTest.testNonExistingOpCode

    This is the third time this failure has popped up. This time it seemed to only impact
tests run using NettyServerCnxnFactory (so it only impacts 3.5 and master) and I was able
to get it to pop up with sufficient frequency when running the tests locally. 
    
    The issue is caused by improper handling of netty's futures. When we call `channel.write(wrappedBuffer(sendBuffer));`
the write is completed asynchronously. The close call `channel.close();` is also asynchronous.
So we can run into the case where the close occurs before a write.
    
    This patch changes our close call to be a callback for the completion of an empty write.
This way we are guaranteed that the channel has "drained" before a close. 
    
    My primary concern with this patch is the channel being used while it is closing (between
the write of an empty buffer and the execution of the close callback). I have added a `closingChannel`
boolean to track that, which I believe is sufficient. Let me know if anyone finds a situation
where that is not the case.

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

    $ git pull https://github.com/afine/zookeeper ZOOKEEPER-2786_third_times_a_charm

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

    https://github.com/apache/zookeeper/pull/327.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 #327
    
----
commit 6e16b20cd2f768e9a027f1d9306ae8126188edb4
Author: Abraham Fine <afine@apache.org>
Date:   2017-08-03T20:02:06Z

    ZOOKEEPER-2786: Flaky test: org.apache.zookeeper.test.ClientTest.testNonExistingOpCode

----


> Flaky test: org.apache.zookeeper.test.ClientTest.testNonExistingOpCode
> ----------------------------------------------------------------------
>
>                 Key: ZOOKEEPER-2786
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2786
>             Project: ZooKeeper
>          Issue Type: Bug
>    Affects Versions: 3.4.10, 3.5.3
>            Reporter: Abraham Fine
>            Assignee: Abraham Fine
>             Fix For: 3.5.4, 3.6.0
>
>
> This test is broken on 3.4 and 3.5, but is broken in "different" ways. Please see the
individual pull requests for detailed descriptions for the issues faced in both branches.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message