New connection behavior with macOS client v1.3.4059.45

Pritunl Pritunl VPN
1

We run our Pritunl VPN server behind Cloudflare to protect the web UI. In the Pritunl config, our public IP is set to the real IP of our server (for the OpenVPN and Wireguard UDP traffic). The sync host is set to our full DNS hostname (Cloudflare). It worked great before our users updated to v1.3.4059.45. Now, all users experience a 40 second delay when connecting. Running tcpdump, the client now tries to connect to the public IP on port 443 and eventually times out. The user eventually can connect, but the long delay is not a great user experience.

I’m not sure for the reason for this change, but it doesn’t seem like whatever the client is trying to do is needed since the connection is eventually allowed.

Is there a setting we can set to avoid this behavior?

Thanks for any help.

2

When the connection code was rewritten the order was inadvertently changed from sorting sync remotes first to sorting profile remotes first. This will be fixed in an update possibly later tonight. Send the service logs from the top right menu to confirm it is only an issue with the remotes ordering.

3

Hi Zach,

Thanks for the help. Here you go (hostname and IP have been obfuscated to vpn.host.com and 1.2.3.4). gist:0b95d8173ce9483e9adff6cd4eaead82 · GitHub

(Too big to post here)

Thanks,

Trevor

4

This has been fixed in the codebase and will be included in the next release.

1 Like