This page will list guidance derived from frequent questions and requests for support from the opnfv-users mail list.

• How to redeploy a virtual Foreman install? I tried clean.sh, But it takes more time but not able to clean up every thing. (original thread)
  • clean.sh may not clean up a virtual deployment (only the Foreman VM). You can cd /tmp/<each node>, and do a "vagrant destroy" to destroy it, where "each node" is going to be controller1..3, compute. This will be fixed in SR1.

• Does Foreman Arno support deployment behind a firewall via proxy?
  • No. There is however, a manual work around documented here

• What is the minimum virtual Foreman Arno install?
  • The minimum install consists of 1 Foreman VM, 1 Controller, and 1 Compute node, with each VM requiring 40GB of space. The Foreman VM and Compute node require 2GB of RAM each minimum, while the Controller requires 4GB. Only a single NIC with internet access is required on the host with a virtual deployment.

• I get the following error: Failed to parse inline template: no implicit conversion from nil to integer at /usr/share/openstack-foreman-installer/puppet/modules/quickstack/manifests/pacemaker/rabbitmq.pp:17, what does this mean?
  • This means there is a conflict between the private IP being handed to the controllers vs. the IPs the controllers should have had. If you check Foreman global paramter controllers_ip_array you will see the IPs the controllers should have. If the IP of your controller is not in that array then that is why you are getting that error. Please check the MAC address you defined in your ksgen_settings.yml file matches the private_mac of the controller in question.

• On a virtual install, my compute node has been created, but it is too small to do anything but run a cirros or two. How can I make it bigger?
  • Before running deploy.sh, add the else/fi section below to where it edits the Vagrantfile and sets the size for controller nodes. It will now do the same for compute nodes. Size can be set to what you think your host machine can handle. In this case it's 16 GB memory and 8 CPUs.

 if [ "$node_type" == "controller" ]; then
   sed -i 's/^.vb.memory =.$/ vb.memory = 4096/' Vagrantfile
 else
   sed -i 's/^.vb.memory =.$/ vb.memory = 16384/' Vagrantfile
   sed -i 's/^.vb.cpus =.$/ vb.cpus = 8/' Vagrantfile
 fi

• One or more of my nova instances do not get a DHCP IP
  • This is a problem where the neutron dhcp agent does not get updated correctly when a new instance spawns. A workaround is to create a neutron router and attach a port to the subnet before you spawn instances or dhcp server.

• I have an instance running, but how to I ssh into it?
  • Guest systems are typically run in their own Network Name Spaces. You will need to reference the name space when invoking ssh. Once you have created a network and added an instance to it, you can find the namespace, and then use it to do the ssh in that namespace. From the controller:

  > ip netns list 
  qdhcp-3890e717-7712-471f-bfca-faf60737b4f5
  
  > ip netns exec qdhcp-3890e717-7712-471f-bfca-faf60737b4f5 ssh cirros@10.0.0.5
  cirros >

• I installed the ISO but could not find the bgs_demo directory.
  • You need to git clone

  git clone -b v1.0 https://github.com/trozet/bgs_vagrant.git