Hi,
We were having the same issue after migrating to Office 365. What seems to fix my issue was adding the default gateway on the tunnel interface (as already mentioned in this thread). As this is not an advised way and also not practical if you have multiple OpenVPN servers. So we needed another solution.
After carefully investigating the problem we found that you can detect the problem with the following PowerShell command
Get-NetConnectionProfile
Name : some-network
InterfaceAlias : Ethernet
InterfaceIndex : 1
NetworkCategory : Private
IPv4Connectivity : Internet
IPv6Connectivity : NoTraffic
Name : my-domain
InterfaceAlias : TAP-Interface
InterfaceIndex : 2
NetworkCategory : DomainAuthenticated
IPv4Connectivity : NoTraffic
IPv6Connectivity : NoTraffic
This command gives you the following values for your OpenVPN TAP-interface IPv4Connectivety status:
- NoTraffic: you are having internet connectivity issues
After some more investigation I was able to solve this issue not by adding a default gateway but something similar an additional default route. This additional default route can be added client side or server side (my preferred option).
Client side you add the following line to your configuration:
route 0.0.0.0 0.0.0.0 vpn_gateway
Server side you add the following line to your configuration:
push "route 0.0.0.0 0.0.0.0"
As you my notice this route will probably never be used as OpenVPN already adds two routes for smaller network segments (0.0.0.0/1 and 128.0.0.0/1) which have a higher priority. But it seems to trick the NLA service in thinking that you are connected to the internet and allowing you to access Office 365.
Hope this helps you all.
Additional references: