[
https://issues.apache.org/jira/browse/AMQ-5238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14933872#comment-14933872
]
Erik Wramner commented on AMQ-5238:
-----------------------------------
This is more important than it seems. When JDBC is used for the messages store
in a master/slave configuration the only current options for scheduled messages
are the in-memory and the KahaDB job scheduler stores. That means that unless
the KahaDB store is highly available there will be serious issues in a failover
situation. Even with a highly available KahaDB store it is not clear if the
broker handles failover correctly when the main store uses JDBC and only the
job scheduler store uses KahaDB. It is quite confusing and not very clear in
the documentation.
Scheduled messages may seem exotic and not a big deal, but ActiveMQ uses them
with the redelivery plugin. That makes them important.
> Add JDBC Persistence for Scheduler
> ----------------------------------
>
> Key: AMQ-5238
> URL: https://issues.apache.org/jira/browse/AMQ-5238
> Project: ActiveMQ
> Issue Type: New Feature
> Components: Job Scheduler
> Affects Versions: 5.9.0
> Environment: JBoss A-MQ 6.0
> Reporter: Jason Shepherd
> Priority: Minor
>
> AMQ-3024 added support for non-KahaDB persistence of scheduled jobs, but a
> JDBC alternative has yet to be added.
> If we're using JDBC persistence adapter, we want all persistence for the
> broker to be done for JDBC. Ideally this would also allows us to view
> historical statistics of the scheduler as per AMQ-5192.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
