[stunnel-users] behaviour when using 127.0.0.1 in the 'connect' field

Tue Mar 10 06:25:33 CET 2020

Hi all - just want to report a problem I solved recently but wanted to get some insights on what was causing the problem.

About me - learnt some unix at university (20 years ago) but nothing too serious.  Recently (1 month ago) acquired own domain name and now poking around the cloud computing / VPS thing.

Project - hand-rolling my own VPN setup on a Ubuntu 18.04 VPS.  OpenVPN is easy since it is a git-clone thing and then just follow the openvpn-install script.  I wanted to add on the Stunnel wrapper because intended to use the VPN in China and apparently their firewall does deep packet inspection and can recognize (and block) openvpn traffic.

Problem - when I set up my stunnel using 127.0.0.1 as the connect destination; it doesn't seem to work (I can see from openvpn window that things seem to pipe through stunnel but then immediately the connection is terminated).  If I replace the 127.0.0.1 with IP of the box I am using (say for example 1.2.3.4); everything works!  The FQDN is ok as well; as long as I don't use 127.0.0.1

Specifically the stunnel.conf:

[OpenVPN]

Accept = 443  # clients connect through 443 to further avoid potential blocking

Connect = 127.0.0.1:1194   #<- this line won't work; but if I replace with 1.2.3.4:1194 then it will work!

Question - My problem is fixed but I am curious if there is any insights on why this is happening given that 1.2.3.4 and 127.0.0.1 are the same machine?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.stunnel.org/pipermail/stunnel-users/attachments/20200310/4e84140e/attachment.htm>