[
https://issues.apache.org/jira/browse/AMQ-4123?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13482628#comment-13482628
]
Christian Posta commented on AMQ-4123:
--------------------------------------
Yah, I hear ya. Im on the fence now too. If we use the KeepAliveInfo, then it
ties the PINGREQ/RESP to the InactivityMonitor. As it is right now, when the
MQTTProtocolConverter sees a PINGREQ, it responds right away with a PINGRESP
regardless of whether there is an InactivityMonitor
(transport.useInactivityMonitor=false).
On the other hand, the InactivityMonitor is the heart-beating on the server
side and the PINGREQ/RESP is the contract. My instincts say to make the
PINGREQ/RESP explicit in the implementation.
idk yet. i'll run my proposed changes by you before I commit.
> Improve MQTT Inactivity Monitoring
> ----------------------------------
>
> Key: AMQ-4123
> URL: https://issues.apache.org/jira/browse/AMQ-4123
> Project: ActiveMQ
> Issue Type: Bug
> Components: MQTT, Transport
> Affects Versions: 5.7.0
> Reporter: Christian Posta
> Priority: Minor
> Fix For: 5.8.0
>
>
> * Keep Alive should be 1.5 * the keep alive value specified by client (the
> grace period described in spec:
> http://public.dhe.ibm.com/software/dw/webservices/ws-mqtt/mqtt-v3r1.html)
> * MQTTInactivityMonitor checking for KeepAliveInfo in the onCommand, but it
> can never get a KeepAliveInfo as it is
> * What to do when user explicitly turns Inactivity Monitor off?
> * Should there be defaults to close potentially dead connections?
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
