[jira] [Resolved] (AMQ-6782) Document about cursorMemoryHighWaterMark is misleading

["Gary Tully (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/AMQ-6782?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Gary Tully resolved AMQ-6782.
-----------------------------
    Resolution: Fixed
      Assignee: Gary Tully

Thanks for the feedback.

I split the description for cursorMemoryHighWaterMark into queue and topic variants.

https://cwiki.apache.org/confluence/display/ACTIVEMQ/Per+Destination+Policies

> Document about cursorMemoryHighWaterMark is misleading
> ------------------------------------------------------
>
>                 Key: AMQ-6782
>                 URL: https://issues.apache.org/jira/browse/AMQ-6782
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Documentation
>    Affects Versions: 5.15.0
>            Reporter: Alvin Lin
>            Assignee: Gary Tully
>
> In http://activemq.apache.org/per-destination-policies.html the description for "cursorMemoryHighWaterMark" is kind of misleading; the description about that the threshold will apply to "the destination's memoryLimit (when defined) " made me think that it applies to TopicSubscription as well. 
> I thought if I configure a topic with memoryLimit of 100MB and system memory limit of 1GB, then TopicSubscription (or more specifically the FilePendingMessageCurosr) uses cursorMemoryHighWaterMark against 100MB. But through testing, code reading, and http://blog.christianposta.com/activemq/activemq-understanding-memory-usage I found out the hard way this is not the case; only queues will use cursorMemoryHighWaterMark against the per destination memory limit. 
> My suggestion is to add additional notes that the configured cursorMemoryHighWaterMark applies only to system memory in the case of topic.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)