version: master branch
page: submodules/oom.git/docs/oom_setup_kubernetes_rancher

These instructions are really good.  However, when I tried to follow them in our WindRiver Project (DMaaP), I found some differences.  Perhaps these apply to other Projects, so they should be added as Notes to the article:

• the exact name for Networks differs between Projects in Titanium GUI.
• It is possible to assign both an internal Network and external Network at the same time using the Launch Instance wizard by selecting both.  However, it seems to be VERY important to select the external network first so it is assigned to NIC1
• I had mixed luck using Ubuntu-18.04 as documented.  It often took several Hard Reboots of the container before I could successfully ssh.  But even then, I found that the config scripts did not execute properly. (there seems to be some consensus of a race condition existing).  So I switched to ubuntu-16.04-cloud-amd64 and had much better success.
• Still, the internal networking between 10.0.0.X IPs never seemed to work.  Perhaps more explanation could be given around the definition of onap_sq_wZ5M used in the example?  I defined a new Security Group based on values from the Security Group in Project Integration-Daily-Release where I assume "things just work", but had no luck in resolving this.  So I had to use the external address for both node values in the cluster.yml file.  i.e both the address attribute and the internal_address attribute were the same.
• Similarly, when invoking the master and slave NFS shell commands, I passed the external addresses