On 11/18/2013 03:17 PM, Sam Jones wrote:
I finally discovered my underlying problem. I had created qpid2 and qpid3 as VMware Workstation clones. Therefore qpid2 and qpid3 had the same broker ID as qpid1. These broker IDs were being saved in the/var/lib/qpidd/.mbrokerdata file. Every time I restarted the qpidd daemons, they were importing this saved value. I stopped the daemons on qpid2 and qpid3, deleted the .mbrokerdata file on both of them, and restarted the daemons. This caused unique broker IDs to be generated for qpid2 and qpid3. Message relaying is working perfectly now.
Thanks for the update, Sam! That sounds like something that others could run into as well so being aware of it is useful.
I've put up a trivial patch for detecting this[1] and logging a warning. A more draconian approach would be to reject connections in such cases. That would make it easier to track down. Thoughts, anybody?
[1] https://reviews.apache.org/r/15757/ --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
