Re: no captive portal page

gusdvg

I'm having trouble with my Captive Portal, which is similar to this. My ps ax | grep light command outputs:

# ps ax | grep light
  709  ??  S      0:00.15 /usr/local/sbin/lighttpd -f /var/etc/lighty-webConfig

I guess my Captive Portal process is not executing. I used to have transparent squid, but have removed it since trying to get Captive Portal working. I'm using 1.2.1-RC1-20080807.

Gertjan

edit : Oops, be carefull - this is a Captive Portal issue - happening on 1.2.1 only.
Yes,

I'm having the same issue.

I was running the 1.2 'stable' for a couple of month now (rock-solid by the way)

FreeBSD pfsense.kyriadfumel.local 6.2-RELEASE-p11 FreeBSD 6.2-RELEASE-p11 #0: Sun Feb 24 16:32:58 EST 2008 sullrich@builder6.pfsense.com:/usr/obj.pfSense/usr/src/sys/pfSense.6 i386

When I updated to 1.2.1 everything went ok, but…
No matter what, the captive portal wasn't answering anymore.
Note that the interface 'Opt1' was up, and PC's connected to my visitors network (a couple of RJ45 jack and WiFi AP's) were servering IP's, given by pfSense.
But No Captive portal anymore.

Shutting down the captive portal made my Opt1 network actif, all clients could connect to the net right a way.

What I figured out already two weeks ago is that the captive portal web server task, lighttpd isn't running on 1.2.1, on 1.2, it is (I use ps -ax ... to detect).

I 'think' that this can't be a driver (or any BSD kernel 7) issue, because the interfaces just work fine, as the DHCP server is (on all my interfaces, except the WAN).

When booting from CD (1.2 stable) everything works as before.

Switching off and on (with the web interface) the captive portal doesn't change anything.

The real question is : Am I the only one that is using the Captive Portal (with Local user manager authentification), what means that something isn't working on my side.
Well, I gues not, now I know that some else has the same problem.

Is there a way to 'clean install' from CD (version 1.2.1) instead of the proposed update/upgrade (which 'seems' to work well).

Throw me any questions you want... and I'll answer.

cmb

First:
1. Please don't hijack other people's threads, the chances that your issue is the same as a 1+ year old thread are pretty slim at best. I split this into a new thread.
2. Please keep all 1.2.1 related discussions in the 1.2.1 board until it's the current stable release.

As to the issue, can you copy/paste your system log and captive portal log?

Gertjan

@cmb:

First:
1. Please don't …

First of all, your right, sorry :-\

@cmb:

As to the issue, can you copy/paste your system log and captive portal log?

Impossible, because 'lighttpd' for the captive portal isn't running.
And, as usual, no errors, or notices, are shown in my log when it concerns 'lighttpd' (from boot to 'going stable' is still 150 lines).

In Status => Services is being marked lighttpd "Captive Portal" as running, but restarting it, or stopping will crash even the 'lighttpd' of the Webconfigurator….. :(

Anyway, I somewhat solved the issue, when running this by hand:

pfsense:/var/log#  /usr/local/sbin/lighttpd -f /var/etc/lightyCaptivePortal.conf
2008-08-11 16:33:14: (plugin.c.165) dlopen() failed : /usr/local/lib/lighttpd/mod_rewrite.so Cannot open "/usr/local/lib/lighttpd/mod_rewrite.so"
2008-08-11 16:33:14: (server.c.621) loading plugins finally failed
pfsense:/var/log#

New in lightyCaptivePortal.conf is the mod_rewrite.so missing LIB error, which isn't present in on the Update/upgrade en ISO install CD.

Extrat of the 1.2.1 lightyCaptivePortal.conf:

## modules to load
server.modules              =   (
                                  "mod_accesslog",
                                  "mod_access", "mod_accesslog",
                                  "mod_fastcgi", "mod_cgi","mod_rewrite"
                                )

When I remove the "mod_rewrite" part, the Captive portal start ok - PfSEnse 1.2.1 is up with captive portal :D

edit : well, euuh, nearly.
When I connect myself with an initial URL like http://www.google.fr/ig?hl=fr then a 404 will be shown.
Making the url look like http://www.google.fr and the login screen will show nicely.

When making a correct login (I'm using the integrated user manager) the text "Redirecting to …" shows up shortly.
But instead of seeing the orginal requested URL (http://www.google.fr), I'm being thrown back to the 192.168.2.1:8000 page = the login screen.
Retyping an URL like www.google.fr then this will bring me now to the Google site, so I'm being authorized, and the captive portal logs confirm this.
So, it works 'nearly'.

I can't say 'bug' here, because I de-activated "mod_rewrite" - this might be related with what I said here (the not being rederected part).

So, it's up to the dev's to decide now - if questions, aks and I'll have a look.

PS : The captive Portal 1.2.1 can't be working for no-one right now : Only 2 'users' noticed this ??

cmb

Thanks for the detailed report, appreciated! You aren't the first or only two, it's been reported by others as well, just haven't had time to look into it. Looks like we aren't getting mod_rewrite, probably related to the build system changes. Should have that fixed shortly.

cmb

Scott just committed a fix for this, the next snapshot built after the time of this post will include it.

Gertjan

I just booted from CD with
pfSense-20080812-0838.iso.gz from 12-Aug-2008 05:32 (1.2.1)

The Captive Portal is up and running.

753 ?? S 0:00.08 /usr/local/sbin/lighttpd -f /var/etc/lighty-webConfigurator.conf
882 ?? S 0:00.06 /usr/local/sbin/lighttpd -f /var/etc/lighty-CaptivePortal.conf

Clients can connect 'as before'.

Great work - thanks :)