Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    Captive Portal fails regularly after upgrading from 2.0.1 to 2.0.2

    Scheduled Pinned Locked Moved Captive Portal
    43 Posts 15 Posters 27.8k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • S Offline
      Slam
      last edited by

      I havent gitsynced the latest updates since the early hours of this morning but so far so good, the original error of captive portal crashing seems to be fixed, as well as some other minor bugs, the redirected lighty error log also reports no problems.

      Big thanks to the devs and pfsensers!

      1 Reply Last reply Reply Quote 0
      • D Offline
        debremarkos
        last edited by

        Hi I am having the same issue, upgraded from 2.0.1 (i386) to 2.0.2 in the services tap captive portal appears down. Interestingly I use radius and openldap for the username and passwords on a separate computer. When a user types in a username and password they get an error saying that page could not be reached, but then if they close the browser and open it they have been authenticated by captive portal and all is working. I am also using esx 4, I use squid (as transparent proxy) and squidguard and a wpad file to auto configure the browser. It all worked fine before the upgrade.

        1 Reply Last reply Reply Quote 0
        • E Offline
          eri--
          last edited by

          Please gitsync to latest changes i think the root cause of this has been fixed now.

          1 Reply Last reply Reply Quote 0
          • D Offline
            debremarkos
            last edited by

            yep did this and after a reboot everything works again, thanks for your help really appreciated.

            1 Reply Last reply Reply Quote 0
            • D Offline
              dgwilson
              last edited by

              @ermal:

              Please gitsync to latest changes i think the root cause of this has been fixed now.

              I updated (sync'd) again. Everything has remained stable over night.

              • David
              1 Reply Last reply Reply Quote 0
              • R Offline
                rhy7s
                last edited by

                I did a gitsync the other day and that fixed the captive portal being stopped issue. What I've noticed since though is that clients have to either renew their IP or turn WiFi Off/On to be redirected to the captive portal after the 1 hour time out. DHCP default and maximum lease times are not set in the configuration so I'm presuming they are 720 seconds and 86400 seconds respectively, both definitely longer than the time out.

                1 Reply Last reply Reply Quote 0
                • W Offline
                  wwifi1944
                  last edited by

                  Same here >:(…upgraded and no more capitve po

                  1 Reply Last reply Reply Quote 0
                  • C Offline
                    cmb
                    last edited by

                    @wwifi1944:

                    Same here >:(…upgraded and no more capitve po

                    gitsync as described here and it will be fixed.

                    1 Reply Last reply Reply Quote 0
                    • M Offline
                      myke
                      last edited by

                      Hi everyone,

                      I confirm with the fix is working.

                      But i have some logs that i don't understand like :

                      lighttpd[34598]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'redirurl' (attacker '172.16.1.37', file '/usr/local/captiveportal/index.php')

                      lighttpd[34598]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'sessionkey' (attacker '172.16.1.141', file '/usr/local/captiveportal/index.php')

                      lighttpd[34598]: (connections.c.137) (warning) close: 73 Connection reset by peer

                      Best regards.
                      Myke.

                      1 Reply Last reply Reply Quote 0
                      • N Offline
                        niebla
                        last edited by

                        Using 2.02 in a production environment so I am reluctant to do a gitsync. I may come in this weekend and give it a try.

                        Interesting that although under service status the captiveportal service says Stopped captive portal is running and users are being asked to login.

                        1 Reply Last reply Reply Quote 0
                        • K Offline
                          kolomalo
                          last edited by

                          Hello.

                          I have the "same issue" Since I upgrade my installation, captive portal don't works as expected. I have had to disable it.

                          My problem is that the ip roules stop working suddenly.

                          f.e. my rule "both any->10.2.0.0/16" stop working and the only way to work around was enable certains IPs on the captive portal (so they have access to internet too, and i don't want this).

                          Thanks!

                          Now i'm gitsyncing.. I'll try after

                          1 Reply Last reply Reply Quote 0
                          • D Offline
                            debremarkos
                            last edited by

                            Hi after gitsyncing and going to 2.0.3 I had a number of other issues, such as under heavy load the firewall blocking everything with nothing entered into the syslog. we have 800 captive portal users and everything was working well in 2.0.1 . I have reverted back to 2.0.1 . The main reason was due to the crashing and the fact that the web gui became really slow, and crashed alot. I run 2 other pfsense devices and I have not gone back to 2.0.2 on them as they do not use captive portal. On all the devices that I have upgraded i have found the performance of the webgiu gets much worse after the upgrade with me having to remove the status widget from the dashboard to make some small improvements.

                            1 Reply Last reply Reply Quote 0
                            • N Offline
                              niebla
                              last edited by

                              Web GUI seems very fast with 2.02. Have not gitsynced yet. Is there a 2.0.3 release? I have not seen it.

                              1 Reply Last reply Reply Quote 0
                              • D Offline
                                debremarkos
                                last edited by

                                when you gitsync you will goto 2.0.3 pre release

                                1 Reply Last reply Reply Quote 0
                                • M Offline
                                  m4st3rc1p0
                                  last edited by

                                  he is right is there any workaround, a lot of error message coming out. Also Captive portal is not working.

                                  @debremarkos:

                                  when you gitsync you will goto 2.0.3 pre release

                                  1 Reply Last reply Reply Quote 0
                                  • E Offline
                                    eri--
                                    last edited by

                                    The latest 2.0.3 is stable from our testing.
                                    Can you try because at the time there were some changes being done.
                                    Now it is marked as stable on our side!

                                    1 Reply Last reply Reply Quote 0
                                    • M Offline
                                      m4st3rc1p0
                                      last edited by

                                      heres the error i got

                                      Jan 31 19:57:20 	lighttpd[21213]: (connections.c.305) SSL: 1 error:140760FC:SSL routines:SSL23_GET_CLIENT_HELLO:unknown protocol
Jan 31 19:57:20 	lighttpd[21213]: (connections.c.305) SSL: 1 error:140760FC:SSL routines:SSL23_GET_CLIENT_HELLO:unknown protocol

                                      @ermal:

                                      The latest 2.0.3 is stable from our testing.
                                      Can you try because at the time there were some changes being done.
                                      Now it is marked as stable on our side!

                                      1 Reply Last reply Reply Quote 0
                                      • H Offline
                                        heper
                                        last edited by

                                        yesterday i've deployed 2.0.3 with a ssl cert from startssl & Radius auth on a Win2K8r2

                                        i've seen that too:

                                        Jan 31 19:57:20 	lighttpd[21213]: (connections.c.305) SSL: 1 error:140760FC:SSL routines:SSL23_GET_CLIENT_HELLO:unknown protocol
Jan 31 19:57:20 	lighttpd[21213]: (connections.c.305) SSL: 1 error:140760FC:SSL routines:SSL23_GET_CLIENT_HELLO:unknown protocol

                                        doesn't seem to affect the portal … i've had +40 portal users all day without complaints

                                        1 Reply Last reply Reply Quote 0
                                        • S Offline
                                          Slam
                                          last edited by

                                          Running latest snapshot of PRERELEASE-2.0.3 (31/01/2013)

                                          I can also confirm what the last 2 posters have reported, though it doesnt seem to affect the CP users.

                                          I am also seeing a lot of the following.

                                          Jan 31 20:23:57 lighttpd[18696]: (connections.c.137) (warning) close: 25 Connection reset by peer
                                          Jan 31 20:23:57 lighttpd[18696]: (connections.c.137) (warning) close: 25 Connection reset by peer

                                          Jan 31 20:02:55 lighttpd[18696]: (request.c.1133) GET/HEAD with content-length -> 400
                                          Jan 31 20:02:55 lighttpd[18696]: (request.c.1133) GET/HEAD with content-length -> 400

                                          Jan 31 19:56:03 lighttpd[18696]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'redirurl' (attacker '10.0.0.109', file '/usr/local/captiveportal/index.php')
                                          Jan 31 19:56:03 lighttpd[18696]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'redirurl' (attacker '10.0.0.109', file '/usr/local/captiveportal/index.php')

                                          Jan 31 17:55:50 lighttpd[18696]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'info_hash' (attacker '10.0.0.78', file '/usr/local/captiveportal/index.php')
                                          Jan 31 17:55:50 lighttpd[18696]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'info_hash' (attacker '10.0.0.78', file '/usr/local/captiveportal/index.php')

                                          Jan 31 17:35:49 lighttpd[18696]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'checklic' (attacker '10.0.0.74', file '/usr/local/captiveportal/index.php')
                                          Jan 31 17:35:49 lighttpd[18696]: (mod_fastcgi.c.2676) FastCGI-stderr: ALERT - ASCII-NUL chars not allowed within request variables - dropped variable 'checklic' (attacker '10.0.0.74', file '/usr/local/captiveportal/index.php')

                                          1 Reply Last reply Reply Quote 0
                                          • E Offline
                                            eri--
                                            last edited by

                                            @heper:

                                            yesterday i've deployed 2.0.3 with a ssl cert from startssl & Radius auth on a Win2K8r2

                                            i've seen that too:

                                            Jan 31 19:57:20 	lighttpd[21213]: (connections.c.305) SSL: 1 error:140760FC:SSL routines:SSL23_GET_CLIENT_HELLO:unknown protocol
Jan 31 19:57:20 	lighttpd[21213]: (connections.c.305) SSL: 1 error:140760FC:SSL routines:SSL23_GET_CLIENT_HELLO:unknown protocol

                                            doesn't seem to affect the portal … i've had +40 portal users all day without complaints

                                            That's just a warning because it tries to forward anything on tcp to port 80 which has been reduced only to tcp traffic on 2.1 version of pfSense.
                                            Since the client is trying an https://www.pfsense.org but the firewall redirection sends it to a simple HTTP talking webserver you get the warning.

                                            1 Reply Last reply Reply Quote 0
                                            • First post
                                              Last post
                                            Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.