I am sorry, I almost missed your mail. I had no idea about this issue before your patch https://review.openstack.org/#/c/12748/. After I read your patch, I have some questions. In your solution, there are two instances of cinder-api, Are there race conditions? Can I start two instances of cinder-api in the same host? 2012/9/12 Clay Gerrard <clay.gerrard at gmail.com> > So we came up with the solution of running a "management" instance of > cinder-api that nova can use to talk to cinder - but isn't exposed to > the user (i.e. it's not in the catalog, and presumably also on > darknet/10. space) > > The expectation is that the volume-actions ext wouldn't be enabled on > the public instance of cinder-api, but nova could call > os-volume_actions on the management instance. > > For us, this requires this patch: > https://review.openstack.org/#/c/12748/ > > -clayg > > _______________________________________________ > OpenStack-dev mailing list > OpenStack-dev at lists.openstack.org > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20120912/3d660741/attachment.html>