[ https://issues.apache.org/activemq/browse/AMQCPP-246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=52350#action_52350 ]
Timothy Bish commented on AMQCPP-246: ------------------------------------- That's not surprising as most TCP sockets won't break when a wire is pulled or a switch goes away. soKeepAlive defines a two hour timeout that will kick the socket but that means that you have to wait two hours. > Failover transport doesn't detect network failures > -------------------------------------------------- > > Key: AMQCPP-246 > URL: https://issues.apache.org/activemq/browse/AMQCPP-246 > Project: ActiveMQ C++ Client > Issue Type: Bug > Components: CMS Impl > Affects Versions: 3.0 > Environment: CMS Consumer running under Linux (Fedora Core 10) > Broker and JMS producer running under Linux (Redhat 7) > Reporter: Daniel Donoso > Assignee: Timothy Bish > Fix For: 3.0.1 > > Attachments: SimpleAsyncConsumer.cpp > > > I tested the CMS SimpleAsyncConsumer example to check the failover transport > funcionality and it doesn't detect the network failures, and after network > restauration the consumer never receives any message. > These are the steps: > 1. The network has a failure. > 2. The Producer sends a message. > 3. After a minute the network is up. (The consumer doesn't detect the > failure) > 4. The consumer doesn't receive the message. > 5. The producer sends other message. > 6. The consumer never receives old and new messages -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.