I have a primary OpenVPN server (site A) hosted on-prem with multiple site-to-site connections; one of these sites is a satellite office (site B). Under normal operation, site B is a site-to-site client of A and exposes it's subnet to site A and one additional subnet/site that only site B has access to via a separate VPN.
I'd like to spin up an OpenVPN server at site B that mirrors the site A server except for the site A <> site B connection to act as a backup if site A goes offline.
All these sites use the embedded OpenVPN frontend on NGFW/untangle
So far, spinning up a server at Site B while the Site A <> Site B connection is live and connecting to Site B's server with new credentials, the VPN connects then drops after <30s. Is this an issue with reusing the VPN's assigned port? What would cause this behavior?
EDIT1: Adding block diagrams for clarity. The solid line are existing, and the dotted lines are proposed. Basically, if site A goes offline, connections to the other sites can be accessed via the site B VPN server. None of the subnets overlap site-to-site, and both servers are using port 1194. Currently checking logs.
