You're correct, you need to be in the libvirtd group on any machine where the controllers are connecting to libvirtd. (I think we had an early prototype where the cloud controller was going to talk to libvirt on the nodes directly, for network management or something. That's obviously a doc bug now.)
We also usually run nova-compute on the cloud controller as well, since the API doesn't generate much load. On Thu, Sep 16, 2010 at 2:19 AM, Manuel Amador (Rudd-O) <[email protected]>wrote: > http://nova.openstack.org/getting.started.html > > It says > > Configuration¶ > ON CLOUD CONTROLLER > > Add yourself to the libvirtd group, log out, and log back in > fix hardcoded ec2 metadata/userdata uri ($IP is the IP of the cloud), and > masqurade all traffic from launched instances > > But unfortunately there is no libvirtd group on my "cloud controller" a > freshly installed ubuntu machine where the steps outlined above for the > cloud controller have been performed. > > Are you sure this is right? It seems to me like one needs to add oneself > to the COMPUTE NODE's libvirtd group, as the CLOUD CONTROLLER does not > actually need to have kvm or will run nova-compute. > > Can you clear this up for me? > > ( Rudd-O @ cloud.com ) > > _______________________________________________ > Mailing list: https://launchpad.net/~nova > Post to : [email protected] > Unsubscribe : https://launchpad.net/~nova > More help : https://help.launchpad.net/ListHelp > >
_______________________________________________ Mailing list: https://launchpad.net/~nova Post to : [email protected] Unsubscribe : https://launchpad.net/~nova More help : https://help.launchpad.net/ListHelp
