Not sure if this is the expected behavior, but this is what happens when testing failover on the activeMQ4-SNAPSHOT (march 23rd). I am using Lingo's async tests in these tests but this is not lingo specific.
Failover works well when starting a client process that attemps to connect to a broker that is not available: once the broker is available, the client connects without problems. But If I stop the broker and there was a client waiting to receive messages, it crashes the client. I did the following steps using the lingo asyn examples with an extenal broker: * I have NO activeMQ broker running * I start a lingo server that connects to brokerURL: failovertcp://localhost:61616)?maxReconnectAttempts=0 * I start a lingo client that connects to the same URL. * Both are waiting for the broker to be available * I start the broker at this URL. * The server receives the message and sends back its result to the client. * The client receives the message and closes. Then, I stop the broker. After a few seconds, the server terminates as well without any thrown exception. Is this the expected behavior? I would have thought that the server would remain active until the broker is restarted. Claude PS: I cross-posted to the old forum before sending here. There does not seem to have much traffic on the old forum so here it is. Original message is at http://forums.activemq.org/posts/list/595.page -- View this message in context: http://www.nabble.com/Failover-failure--t1338117.html#a3576962 Sent from the ActiveMQ - User forum at Nabble.com.
