• 25 Posts
  • 162 Comments
Joined 1 year ago
cake
Cake day: August 15th, 2023

help-circle


  • So…both cloudflare tunnels and tailscale funnels would require to have the service exposed to have a link with a tailscale-running server, right? As in, every single of my services/containers running via funnel would need their own tailscale install. If it’s a web service I can publish it via the nginx proxy-tailscale one, but if it’s just a service with no defined URL, just the port, I need a tailscale funnel in every single container of these, right? Or am I getting this wrong?

    Thanks!




  • …what support? They barely reply any queries people post in their google groups. If you go there you’ll see most people will try to reach them either due to servers down (the main issue at hand) or login issues which never get fixed (the longest standing issue, better create a different new subdomain) from what I’ve seen. I’ve also tried repeatedly to reach them regarding changing the token access, but with no luck. It’s a free service so I can’t complain, but the only support you actually will get is from other users, and for some scenarios that’s not quite enough.

    EDIT: Oh wow right after posting this I just saw they actually replied regarding the SSO/tokens issue most people have (SSO failed due to the reddit snafu, you end up with just the token and no chance to do any further changes to your account again). This has been an ongoing issue for over two years, I just saw they finally replied (I think for the first time) a couple of weeks ago.



  • Does it matter?

    No, it does not change, but why is this something of concern? The problem is duckdns DOES NOT REPLY providing DNS replies, not to my own servers, but to people outside looking for my servers by typing their address. Duck fails to provide a response to those queries, and users get timeouts. I can frequently replicate this with either dig or nslookup, from different machines, either inside my network or at random connections.

    I managed today to run certbot to register 2 new subdomains that yesterday consistently failed with a long timeout during THE WHOLE DAY. Today the same certbot command on the same server ran straight at the first attempt.

    So…yeah. Unreliable.