[jboss-jira] [JBoss JIRA] (WFLY-13007) New subsystem for configuring clustering API implementations

[Paul Ferraro (Jira)]

     [ https://issues.redhat.com/browse/WFLY-13007?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Paul Ferraro updated WFLY-13007:
--------------------------------
    Description: 
Currently, WildFly creates services for the following objects for every JGroups channel:
* CommandDispatcherFactory
* Group

... and creates services the following objects for every Infinispan cache:
* SingletonServiceConfiguratorFactory
* ServiceProviderRegistry
* Registry
* Group

Only some of these services are ever started.  If we instead created a separate subsystem that configures these objects, we could significantly reduce the number of services created (but never started) by the jgroups and infinispan subsystems.

This has a number of benefits for users as well:
* Increases overall visibility of the public clustering API
* Simplifies the current mysterious state of service wiring by establishing explicit capability requirements where necessary
* Simplifies inter-component capability requirements (e.g. SingletonServiceConfiguratorFactory depends on ServiceProviderRegistry and CommandDispatcherFactory).

  was:
Currently, WildFly creates services for the following objects for every JGroups channel:
* CommandDispatcherFactory
* Group

... and creates services the following objects for every Infinispan cache:
* SingletonServiceConfiguratorFactory
* ServiceProviderRegistry
* Registry
* Group

Only some of these services are ever started.  If we instead created a separate subsystem that configures these objects, we could significantly reduce the number of services created (but never started) by the jgroups and infinispan subsystems.

This has a number of benefits for users as well:
* Increases overall visibility of the public clustering API
* Simplifies the current mysterious state of service wiring by establishing explicit capability requirements where necessary



> New subsystem for configuring clustering API implementations
> ------------------------------------------------------------
>
>                 Key: WFLY-13007
>                 URL: https://issues.redhat.com/browse/WFLY-13007
>             Project: WildFly
>          Issue Type: Feature Request
>          Components: Clustering
>    Affects Versions: 19.0.0.Beta1
>            Reporter: Paul Ferraro
>            Assignee: Paul Ferraro
>            Priority: Major
>
> Currently, WildFly creates services for the following objects for every JGroups channel:
> * CommandDispatcherFactory
> * Group
> ... and creates services the following objects for every Infinispan cache:
> * SingletonServiceConfiguratorFactory
> * ServiceProviderRegistry
> * Registry
> * Group
> Only some of these services are ever started.  If we instead created a separate subsystem that configures these objects, we could significantly reduce the number of services created (but never started) by the jgroups and infinispan subsystems.
> This has a number of benefits for users as well:
> * Increases overall visibility of the public clustering API
> * Simplifies the current mysterious state of service wiring by establishing explicit capability requirements where necessary
> * Simplifies inter-component capability requirements (e.g. SingletonServiceConfiguratorFactory depends on ServiceProviderRegistry and CommandDispatcherFactory).



--
This message was sent by Atlassian Jira
(v7.13.8#713008)