Devs here need to connect to on-prem client VPNs (Cisco, Fortinet, etc.) and our NordLayer instance simultaneously. Routing tables get mangled. It's a mess.
How are you solving this without having them reboot every time they switch? Some of us have to run both to hit our data warehouse. Tried split-tunneling configs, but it's brittle.
SQL is enough
Don't run both simultaneously on the same OS routing table. It's asking for trouble and leaks.
Give them a dedicated VM or container for the client VPN. Route all client traffic through it.
Keep NordLayer on the host for everything else.
If they need both for the warehouse, that's a network design issue, not a dev machine config.
null
Agree with the VM approach - that's how we survived the Great Salesforce-to-Hubspot Migration of '21. The caveat is the latency hit for devs who need to interact with the client VPN resources in real time, like querying their live databases.
If you go the VM route, bake the VPN config into the image. Otherwise you're just shifting the mangled routing table problem into a virtual machine and creating a support ticket two months from now when the dev's DHCP lease expires.
It's still a band-aid, though. The real fix is getting the client to whitelist your NordLayer egress IPs. Good luck with that procurement conversation.
Expect the unexpected