[
https://issues.apache.org/jira/browse/AMQ-5380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14156503#comment-14156503
]
Andreas Gies commented on AMQ-5380:
-----------------------------------
When keepDurableSubsActive=false is set on the Broker, the reconnected
subscriber is configured correctly.
> Prefetch size for reconnecting Durable Subscribers
> --------------------------------------------------
>
> Key: AMQ-5380
> URL: https://issues.apache.org/jira/browse/AMQ-5380
> Project: ActiveMQ
> Issue Type: Bug
> Affects Versions: 5.10.0
> Environment: All
> Reporter: Andreas Gies
> Priority: Minor
> Attachments: PrefetchSizeSpec.scala
>
>
> I am using Active MQ 5.10 within a Karaf based application and have spent the
> last couple of days investigating an OOM Exception we were getting
> occasionally.
> It turned out, that the OOM happens occasionallly when we reconnect a durable
> subscriber that would have a large amount of data in it even though we have
> specified a rather small prefetch size for durable subscribers (5) within the
> broker configuration.
> Further analysis showed that after closing and reconnecting the same
> subscriber it has a prefetch size of 1000 (the default value). That would
> cause an OOM if those 1000 messages exceed the JVM heap size.
> Specifying the prefetch size within the broker url from the client side works
> fine, but we would like to keep the setiing on the broker side as not all
> clients are our development responsibility.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
