[jboss-user] [JBoss Messaging] - JBM clustering scenario

http://www.jboss.org/index.html?module=bb&op=viewtopic&t=149834 This is related to another posting (http://www.jboss.org/index.html?module=bb&op=viewtopic&t=149834) - but with a slightly simpler setup. I have two clustered JBoss instances (JBoss1 and JBoss2) using a single MySQL database as the message store (InnoDB engine) client1 sends messages in - connects to Jboss1 (Node 0) client2 receives messages - connects to Jboss2 (Node 1) I send in 40000 messages. During the send - client1 and/or client2 display the following message.. | 11:08:13,730 INFO @Thread-10 failureDetected(FailoverCommandCenter.java:117) [FailoverCommandCenter | ] JBoss Messaging server failure detected - waiting for failover to complete... | The client then waits 60s and continues. It does not appear that the clients have moved to another JMS connection (on every send/receive I print out the connection serverID and it doesn't change). After the test completes - there are messages still in the database associated with Node 1 (i.e. channel ID in JBM_MSG_REF, corresponds to Node 1 in JBM_POSTOFFICE). The receiver program will not read them. Now, if I stop JBoss2 - the messages failover to Node 0 Then, re-start JBoss2 (Node 1 channel_id is refreshed in the database) and run client2 - the receiver - it still doesn't get them. Now stop JBoss1 - the messages fail-over to Node 1. Run the receiver (client2) now - it will pick up the messages OK... 1. To get to the source of the problem - what will trigger the initial exception? How much of a network glitch and how can we tune this? 2. Please can someone explain why it appears that fail-over has to happen in order to release the messages? BTW, I have run the example programs for a distributed queue and it appears to work OK. Any help/suggestions appreciated. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4210345#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...