Em agosto 14, 2020 3:58 Riccardo Paolo Bestetti via arch-general escreveu:

After a reboot, the first openvpn-client@ instance I try to start takes almost exactly two minutes to start. The instances before that one start just fine in a few seconds.

Guess you meant: "The instances *after* ..."

When that happens, I can see from journalctl that the client actually starts in the first few seconds after the systemctl command. But then, the command doesn't terminate for two more minutes (with no further journal entries).

Openvpn has quite good logging capabilities that you can put to use here.

Has anyone seen this before? What could it be?

Without knowing more, my first guess is that you still don't have connectivity when that first openvpn client starts. 2 minutes matches exactly the 120 seconds default ping-restart parameter. So, what happens is, the client starts, you have no connectivity then, after two minutes, ping-restart kicks in, and your connection gets through. So, get a network manager that can properly trigger network-online.target. Or, if your network manager is triggering it, then it means your network is not quite ready when it does. Regards, Giancarlo Razzolini