cayenne-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrus Adamchik (JIRA)" <j...@apache.org>
Subject [jira] [Closed] (CAY-2043) ServerRuntimeBuilder: use DataDomain name for the default DataNode
Date Mon, 28 Dec 2015 14:35:49 GMT

     [ https://issues.apache.org/jira/browse/CAY-2043?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Andrus Adamchik closed CAY-2043.
--------------------------------
       Resolution: Fixed
    Fix Version/s: 4.0.M3

> ServerRuntimeBuilder: use DataDomain name for the default DataNode
> ------------------------------------------------------------------
>
>                 Key: CAY-2043
>                 URL: https://issues.apache.org/jira/browse/CAY-2043
>             Project: Cayenne
>          Issue Type: Improvement
>    Affects Versions: 4.0.M2
>            Reporter: Andrus Adamchik
>            Assignee: Andrus Adamchik
>            Priority: Minor
>             Fix For: 4.0.M3
>
>
> When a DataNode is created implicitly by ServerRuntimeBuilder, it is assigned a hardcoded
name of "cayenne". Turned out this can be a problem sometimes. 
> E.g. in LinkMove (a data copy framework at https://github.com/nhl/link-move) a separate
Cayenne stack is used for each RDBMS source, and then one extra stack is used for the data
target. If we are using default configs for both source and target, source stack binds its
Transaction to the processing thread, and then source connections leak into target operations
because the corresponding DataNode names match. There are probably other ways to fix it (LinkMove
should split producer and consumer into separate threads), but still a good idea to not hardcode
the DataNode name.



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

Mime
View raw message