[ 
https://issues.apache.org/jira/browse/GEODE-8498?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17202205#comment-17202205
 ] 

ASF subversion and git services commented on GEODE-8498:
--------------------------------------------------------

Commit 22f2c5247411f646521f47f6c0f900feaa3aa65c in geode's branch 
refs/heads/develop from John Hutchison
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=22f2c52 ]

GEODE-8498: make AbstractSubscription write to channel synchronously (#5550)

Co-authored-by: Ray Ingles <ring...@vmware.com>
Co-authored-by: Darrel Schneider <dar...@vmware.com>
Co-authored-by: Jens Deppe <jde...@vmware.com>
Co-authored-by: Sarah Abbey <sab...@vmware.com>

> Redis messages written to Netty channel sometimes delivered out of order
> ------------------------------------------------------------------------
>
>                 Key: GEODE-8498
>                 URL: https://issues.apache.org/jira/browse/GEODE-8498
>             Project: Geode
>          Issue Type: Bug
>          Components: redis
>            Reporter: Sarah Abbey
>            Assignee: Sarah Abbey
>            Priority: Minor
>              Labels: pull-request-available
>
> {code:java}
> [err]: Mix SUBSCRIBE and PSUBSCRIBE in tests/unit/pubsub.tcl
> Expected 'pmessage foo.* foo.bar hello' to be equal to 'message foo.bar hello'
> *** [err]: Mix SUBSCRIBE and PSUBSCRIBE in tests/unit/pubsub.tcl
> Expected 'pmessage foo.* foo.bar hello' to be equal to 'message foo.bar hello'
> {code}
> When writing responses to a Netty channel, the messages may get written out 
> of order in the `writeAndFlush` call.  To verify this bug, we used a single 
> thread to call `writeAndFlush` and added logging right before calling 
> `writeAndFlush` and after `afterWrite` in the callback listener.  We saw the 
> messages get logged in the proper order before `writeAndFlush` was called, 
> but the messages were logged out of order in the callback listener and 
> arrived in the wrong order on the client.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to