Re: Cannot restart a failed cluster node (catch-up connection closed prematurely)

[Jaakko Nyman]

I have added the following jira issue with Java test code to repeat the problem: https://issues.apache.org/jira/browse/QPID-3495 I tried to be as thorough as possible but i'll try to help if there is anything additional you need. 19.09.2011 16:50, Alan Conway kirjoitti: On 09/19/2011 07:36 AM, fijany wrote: In addition: - this has been tested on both qpid 0.12 and the mrg 1.0 with the same results - we are using QPid Java JMS client - restart seems to work if no messages are produced/sent during the node downtime Does this mean that all of the producers would need to be stopped during the time nodes are being restarted on a cluster? This would be quite impractical for us. You shouldn't need to stop producers for failover to work. Please create a JIRA for this, with detailed instructions on how to reproduce. You can assign it to me. Thanks, Alan. -- Jaakko Nyman | Software Designer Mobile: +358407088788 EmblaCom Oy P.O. Box 169 | FI-00211 | Helsinki | Finland Visiting address Vattuniemenkuja 5 Main: +358 424 10101