You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by "Rick Kellogg (JIRA)" <ji...@apache.org> on 2015/10/09 02:30:29 UTC

[jira] [Updated] (STORM-179) integration testing of custom serializer code without running a multi-node cluster

     [ https://issues.apache.org/jira/browse/STORM-179?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Rick Kellogg updated STORM-179:
-------------------------------
    Component/s: storm-core

> integration testing of custom serializer code without running a multi-node cluster 
> -----------------------------------------------------------------------------------
>
>                 Key: STORM-179
>                 URL: https://issues.apache.org/jira/browse/STORM-179
>             Project: Apache Storm
>          Issue Type: New Feature
>          Components: storm-core
>         Environment: ubuntu linux 
>            Reporter: Chris Bedford
>            Priority: Minor
>
> I need to integration test  various customer serializers we are writing for our current  project.
> As I understand it, only if a spout or bolt is sending a tuple to a tuple in another JVM does kryo serialization occur.  That's a good performance optimziation, but it would be nice to be able to disable it for testing purposes.
> With such an option one could  force that custom  serialization code to be invoked  even when  running everything in
> memory  in one JVM   (via LocalCluster).  That makes for more convenient debugging and faster turn-around.
> thanks
>  - chris



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)