Guest Access via Captive Portal - problem with page not always appearing

[cenders]

I have set up a guest wifi network with an access point on an ethernet port of the PAN firewall.  The guest wifi network is unsecured.  I am using a web-form to ensure the guests see our logo, terms, and type in the correct password to proceed.

The setup works most of the time.  On occasion when using this setup via an Apple iOS device (iPhone, iPad) I don't get the portal page and instead receive an error on the idevice, "Hotspot login cannot open the page because the network connection was lost".  I can then only press cancel and it disconnects me from that wifi network.  If I hit cancel and try again, the same problem appears.  When hitting cancel on this the second time, I get presented with "use without internet or use other network".  Now if I click use without internet, and then open Safari and a website I do indeed get the captive portal.

I do not have this problem on any non iOS device (windows laptop, etc).

So, why sometimes does this happen to me?  It is not always repeatable on an iOS device.  I believe this is because the apple device is trying to initially open an HTTPS page.  Does that require me to enable SSL decryption?

[cenders]

Here is a capture of my security rules regarding the guest network, in case that helps.

[cenders]

For testing purposes, does anyone know if I can configure the captive portal to be HTTP instead of HTTPS?

[Kenyon]

I had a similiar issue to where the Captive Portal would display some of the time and I would get the generic browser error page at other times.  Long story short, the Captive Portal interface has a config setting for a threshold as to the number of attempts an IP can make to it within a certain period of time.  For me, it was 1 request every 2 seconds.  We upped this to something like 3 requests every second and the problem went away.

The problem in my case was that when a system came on the network and the user begins typing in their search parameter in the browser bar, the browser would perform a kind of prefetch function for the content as they typed it in.  These requests were being sent to the CP and causing it to hit its threshhold.  Sebsequently, by the time the user presses enter, they would get the generic browser error page about the resource not being able to be reached instead of the CP due to the CP dropping the request.

Hope this helps...someone.