Using deploy.sh and adapted base-config from https://github.com/trozet/bgs_vagrant gives follwong errors:
./deploy.sh.orig -baseconfig /root/opnfvksgen_settings.yml ... The box 'chef/centos-7.0' could not be found....
By changing chef/centos-7.0 to opnfv/centos-7.0 in the script, deploy.sh continues until:
Detecting network configuration... Network detected: multinetwork! default X.X.X.X 0.0.0.0 UG 0 0 0 enp3s0f0 Default Gateway Detected Default Gateway: X.X.X.X Default Gateway Detected on Public Interface! Will setup NAT from Admin -> Public Network on VM! Setting node gateway to be VM Admin IP Gathering network parameters for Target System...this may take a few minutes horizonpublic_vip: Neutron Floating IP range: X.X.X.X to X.X.X.Y Foreman VM is up! Parameters Complete. Settings have been set for Foreman. Starting Vagrant! Bringing machine 'default' up with 'virtualbox' provider... ==> default: Box 'chef/centos-7.0' could not be found. Attempting to find and install... default: Box Provider: virtualbox default: Box Version: >= 0 The box 'chef/centos-7.0' could not be found
Seems the value chef/centos-7.0 is hardcoded somewhere.
BTW: I am pretty familiar with Openstack, Openstack Neutron, Foreman, Opendaylight and also Ceph. In my opinion I need a lot more to configure than just IPMI/private/admin MACS and IPs of my nodes. What about Neutron floating ranges? What about my public range doesn not have a /24 mask ? I have actually 4 separate 1G and 2 separate 10G NICS in every node , and the IPMI network is completely dedicated. Makes 7 networks. Is it also possible to use bonding (mode=6) for link failure and bigger multi-connection bandwith? Not all my nodes have the same interface orders as well. How I make sure the two 10G NICS are used as data networks?
Would be nice to get an OPNFV testbed running here.
Any help appreciated.