Skip to main content
naltakeb
naltakeb
New Member
March 11, 2022
Question

Captive Portal issue for Users Over Site to Site IPSec VPN

  • March 11, 2022
  • 2 replies
  • 2681 views

 

branch user --> 80F FW ==> IPSEC VPN ==> 1100E FW ==> Captive Portal ==> Internet

 

Common issue with the example scenario:

A common issue when configuring Captive Portal for this sample scenario is that Captive Portal does not work perfectly for vpn users behind 80F FW that come over the site-to-site IPSec vpn to the 1100E FW in order to access the internet. Common symptoms are:

  • Site-to-site vpn is up and working
  • Captive Portal works perfectly for local users behind 1100E FW
  • Captive Portal works for local users behind 80F FW but some image on Web Captive Portal do not appear perfectly 

2 replies

sharmaj
Staff
sharmaj
Staff
March 11, 2022

Hello,

 

Regarding this:

  • Captive Portal works for local users behind 80F FW but some images on Web Captive Portal do not appear perfectly 

Is it only about image not being appearing perfectly or the user does not get to authenticate as well?

 

If it is about imperfection in the image, please try to inspect by right-clicking on the browser and check under the network tab, reload the page and check if you see any error per se.

xsilver_FTNT
Staff
xsilver_FTNT
Staff
April 7, 2022

Personally I'm not a big fanboy of a Captive Portals, unless they are there for some reason like Disclaimer page, or email collector on FGT. Especially as they usually affect all the traffic passing through interface and handling exceptions is a bit painful. So I prefer per-policy identity and authentication handling. Which allows me to use stuff like FSSO, auth session inheritance for eg. from VPN so user do not need to authenticate multiple times, etc.

Another thing is that Captive portals are usually on ingress side (as below) not on egress.
Users - PC - Captive portal - interface - FGT...

 

Not sure I'd clearly connect missing images to captive portal. How about to have some more solid proof via debug, at least something like 'flow debug'. Which could be even filtered to specific site or image source, in case the issue is at least somehow reproducible and not completely random and intermittent.

If you have some evidence, then I'd suggest to open TAC ticket on it.

 

Terms & ConditionsAccessibility statement