[
https://issues.apache.org/activemq/browse/AMQ-2196?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=57314#action_57314
]
Timothy Bish commented on AMQ-2196:
-----------------------------------
The current version of ActiveMQ-CPP will send the inactivity durations
correctly.
> InactivityMonitor Issues
> ------------------------
>
> Key: AMQ-2196
> URL: https://issues.apache.org/activemq/browse/AMQ-2196
> Project: ActiveMQ
> Issue Type: Bug
> Reporter: Marc Breslow
> Assignee: Timothy Bish
> Fix For: 5.4.0
>
> Original Estimate: 1 hour
> Remaining Estimate: 1 hour
>
> InactivityMonitor should not simply use the minimum of the local and remote
> maxInactivityDuration parameter
> InactivityMonitor.java uses the following configuration
> readCheckTime =
> Math.min(localWireFormatInfo.getMaxInactivityDuration(),
> remoteWireFormatInfo.getMaxInactivityDuration());
> initialDelayTime =
> Math.min(localWireFormatInfo.getMaxInactivityDurationInitalDelay(),
> remoteWireFormatInfo.getMaxInactivityDurationInitalDelay());
> Because of this, I can't simply define a maxInactivityDuration on my JMS
> server. I need to make sure that the clients specify a value that is greater
> then or equal to the server configured value.
> CMS doesn't yet supply this parameter when it makes an openwire connection so
> in my mixed Java/C++ application environment, I can't bump the inactivity
> timeout > 30s (the default). Why should the server setting for how long to
> go before timing out a connection be influenced by what the client requests?
> Client can always terminate the connection.
> I suggest that this be changed to
> readCheckTime = localWireFormatInfo.getMaxInactivityDuration();
> initialDelayTime =
> localWireFormatInfo.getMaxInactivityDurationInitalDelay();
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
