Found in a Microsoft forum
Well, I have located the problem. I am posting the solution here in case anyone else has this issue. It has to do with the cluster discovery between the cluster nodes, NOT registry issue.
The “Remote Registry” service got disabled somehow on cluster node 2.
During the setup it detects that it is a clustered instance so it queries the cluster information. It queries all of the nodes that are candidates to host SQL. It scans itself. Then it uses the “Remote Registry” service on the other nodes to get information it needs to proceed.
In our case Node 2 had that service disabled, Node 1 had it enabled(Manual, but still able to be launched on demand). When node 2 needed system information for node 2 it accessed the registry directly to get the information. It used remote registry to query node 1. This is the first patch that worked.
During the node 1 install when it needed information from node 1 it looked locally to the registry. When it tried to get information via remote registry from node 2 (where it is disabled, which prevents it from starting on demand) it fails.
The Detail.txt log had a generic “Network path not found” error. I verified that DNS and IP were correct and communicating correctly. I grabbed the 0x Error code from that section of the log and found the remote registry to be the issue. The forums I found talked about it running with the wrong credentials(NetworkService vs LocalService) but when I went to verify the service accounts I noticed that it was disabled on Node 2.
Just remember it isn’t the node you are installing on that has the issue, it is the remote node.