My enviroment two machines: On one machine(node1), cinder-api cinder-volume cinder-scheduler and qpidd are running On other one(node2), only cinder-volume running and cinder-volume use the same qpid with node1. Normally, create/delete volume can work well. However, I encountered a problem: when I stop cinder-volume service(kill cinder-volume process), the cinder-volume queue on qpid will be removed. Anyone encoutered the similar problem? And how to solve it? -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack/attachments/20130521/83a91b83/attachment.html>