Hi, this is just in time for this bug's 10 years anniversary. Dear Utopia Team, how about reacting to the bug report before the ten year cake is baked?
On Tue, Jul 26, 2016 at 03:59:27PM +0200, J. R. Schmid wrote: > in a OpenVPN setup where everything is proven to work as expected using the > `openvpn` command line client, the following happens when using > NetworkManager > to connect to the VPN instead: > > - first, the connection is established successfully > - but then, no extra routes or DNS servers are set I am using OpenVPN 2.7, NetworkManager and nm-applet (in KDE Plasma) on Current Debian Unstable in June 2026. The server is running Debian trixie, OpenPVN 2.6, and systemd-networkd. The server is configured to push IPv6 nameservers and the two routes necessary to reach them:: push "route-ipv6 2a01:238:43fa:bc81::/64" push "route-ipv6 2a01:238:43fa:bc1e::/64" push "dns server 51 address 2a01:238:db8:bc81::35:100 2a01:db8:43fa:bc8e::35:100" push "dns server 51 resolve-domains ka51.example.com" While the two routes show up in the client's routing table, the DNS servers and the resolve domain do not show up in systemd-resolved. When I manually configure the DNS servers in the OpenVPN connection inside nm-applet, systemd-resolved is updated correctly. This is both evidence that the problem most probably sits between network-manager and OpenVPN and proof that network-manager can properly talk and configure systemd-networkd. I am administrator on the server, and the server is not SO important that I can't do experimental configurations there. Let me know what I can do. Also, I'd like to say that OpenVPN upstream is rather helpful and willing to talk to cooperating packages like network-manager to deliver information beyond the offical docs, hunt bugs and to give advice. One just needs to talk to them. Greetings Marc
