| View previous topic :: View next topic |
| Author |
Message |
fronzo User

Joined: 23 Apr 2009 Posts: 19
|
Posted: Thu May 07, 2009 9:33 pm Post subject: CoovaChilli on OrangeMesh or MeshConnect |
|
|
Hi everybody,
I was wondering if OrangeMesh or MeshConnect integrate any option to be used to enable ChilliSpot in Robin, in order to redirect users to a captive portal and allowing to identify them against a radius server.
In other words I'm looking for a setting as the "captive portal providers / other radius provider" in the Open Mesh Dashboard.
If such a feature is unavailable within OrangeMesh or MeshConnect, is there any way to enable chilli manually?
Thanks! |
|
| Back to top |
|
 |
ispyisail Site Admin


Joined: 12 Sep 2008 Posts: 2858 Location: New Zealand
|
Posted: Fri May 08, 2009 4:26 am Post subject: |
|
|
I too want to know the answer to this question!
This is all i could find
Only registered users can see links on this forum! Register or Login on forum! |
|
|
| Back to top |
|
 |
funchip Moderator

Joined: 12 Sep 2008 Posts: 253 Location: Sacramento, California, USA
|
Posted: Sat May 09, 2009 1:07 am Post subject: |
|
|
Yes, the feature is available on MeshConnect. To test CoovaChilli setting, go to a demo site
Only registered users can see links on this forum! Register or Login on forum! |
The dashboard has a migration feature that allows you to import your network from open-mesh network easily. So use that function for your convenience...
By the way, to redirect your nodes from open-mesh to the demo dashboard, make sure you turn on the SSL option on the MeshConnect setting, and set the alternative dashboard check-in URL on open-mesh as below
| Code: | | ssl.perfora.net/wifimesh.trigmax.com/meshconnect |
The demo site is for "demo" only, as we are running test and making changes without notice. So please don't use the site as a production service. We expect you to download the package and run the dashboard on your own server.
Let me know if you have any problem. |
|
| Back to top |
|
 |
fronzo User

Joined: 23 Apr 2009 Posts: 19
|
Posted: Sat May 09, 2009 10:16 am Post subject: |
|
|
Thanks a lot funchip.
I had already found the latest version of MeshConnect and installed it on a test server. It works great! I still haven't tested the captive portal redirection but I have no doubt it works as expected.
Instead of forwarding to the new dashboard from open-mesh I preferred to update my nodes manually adding the new dashboard on the DFE file. With such a configuration I didn't need to enable SSL on my test server.
The dashboard is really nice. There are only a couple of features that I feel really useful in the Open Mesh dashboard which don't seem to have been implemented in MeshConnect: the graph showing the bandwidht usage over time, the list of the connected users and, above all, the network diagram showing the paths used by the nodes to reach the gateways.
Great work anyway. Keep up with it!
Thanks very much! |
|
| Back to top |
|
 |
foxtroop11 Moderator

Joined: 22 Mar 2009 Posts: 763 Location: Ansbach, Germany and sometimes the States
|
Posted: Tue May 12, 2009 9:21 pm Post subject: |
|
|
I had some time so I gave this meshconnect a go. The fact it allows coova settings is a huge plus, what I'm having an issue with is every 10 minutes on the dot I'm getting logged out of a chillispot system and thrown back to the splashpage. I'm confident this is an issue with the dashboard because the same routers with the same captive portal settings on Open-Mesh dashboard is not having the issues.
Any thoughts? |
|
| Back to top |
|
 |
foxtroop11 Moderator

Joined: 22 Mar 2009 Posts: 763 Location: Ansbach, Germany and sometimes the States
|
Posted: Thu May 14, 2009 6:46 am Post subject: |
|
|
| Spent alittle more time with it. The meshconnect dashboard is setting something in coova that's causing the inactivity attribute to work incorrect. If I bump up my inactivity value in my billing back end I can up the time it happens. If I delete the entry all together it falls back to the default 10 mins. I guess I have to figure out how to edit the meshconnect php files to setup coova correct. |
|
| Back to top |
|
 |
funchip Moderator

Joined: 12 Sep 2008 Posts: 253 Location: Sacramento, California, USA
|
Posted: Thu May 14, 2009 4:36 pm Post subject: |
|
|
Can you SSH to the node and capture two files when the Coova operaton failed?
/etc/update/received.clean
and
etc/update/update.arg
The files should be updated every 5 minutes.
Dashboard mainly passes the settings to the node. There is a possibility that the settings were passed over incorrectly or incompletely. |
|
| Back to top |
|
 |
lucianoinacio User

Joined: 30 Mar 2009 Posts: 12 Location: Brasil
|
Posted: Fri May 15, 2009 4:15 pm Post subject: |
|
|
i use chilli with radius.
to cancel user i do it in radius.
In meshconnect how can i disconnetc the user?? How can i do it in dashboard??
Any table in msql?? any downloaeded file??
Any sugestion??
Luciano Inacio |
|
| Back to top |
|
 |
funchip Moderator

Joined: 12 Sep 2008 Posts: 253 Location: Sacramento, California, USA
|
Posted: Sun May 17, 2009 11:40 pm Post subject: |
|
|
On MeshConnect, the client access control is done by the captive portal daemon (nodogsplash, coovachilli, etc). I believe chilli daemon should be the one to block the client's access in your case. Dashboard is doing nothing more than passing the admin setting to the ROBIN firmware, which in turn sets the operation of chilli daemon.
I'm new to the chilli world, too. So I'm also listening...  |
|
| Back to top |
|
 |
foxtroop11 Moderator

Joined: 22 Mar 2009 Posts: 763 Location: Ansbach, Germany and sometimes the States
|
Posted: Fri May 22, 2009 6:12 am Post subject: |
|
|
I took a look at those files you told me to examine. Everything appears normal but I've spoke with another user that is seeing the same thing. It looks as if coova is restaring. I can't be certain yet but something is for sure wrong when I switch from Open-Mesh to Meshconnect. Everything works just fine with OM, but the moment I switch over it starts to act up. By default my captive portal/free-radius setup has a 10 minute time out for inactivity. I've timed it and almost exactly every 10 minutes I get the splash page/login using Meshconnect even though I'm not inactive. I then went and raised the 10 minute variable for the user in mysql to 20 minutes and I'm certain it went almost 20 minutes.
I need to do further testing, but I'm wondering if Open-Mesh is setting something that MeshConnect is either not changing or not seeing. When the nodes switch from OM to MC, does the entire coova config get altered. I've looked into the code and it appears only certain things are changing. I guess I should do a comparision of OM settings to MC.
Is anyone else actually using MC with a captive portal backend? I'm also trying to bring back the map function since most of the code is still there. |
|
| Back to top |
|
 |
funchip Moderator

Joined: 12 Sep 2008 Posts: 253 Location: Sacramento, California, USA
|
Posted: Fri May 22, 2009 6:38 am Post subject: |
|
|
@foxtroop11, frankly, I don't know the answer rigth off my head. Now, if willing, we can do the debugging collaboratively. Can you do the following things for me:
1. Set up a temporary account for me on your radius/UAM server.
2. Show me your setting on both open-mesh and meshconnect dashboards.
3. Allow me to run the test on your radius/UAM server using the temporary account and duplicate the case.
You may want to send me the setting in a private message. I'll do a test and post the finding afterward. |
|
| Back to top |
|
 |
foxtroop11 Moderator

Joined: 22 Mar 2009 Posts: 763 Location: Ansbach, Germany and sometimes the States
|
Posted: Fri May 22, 2009 7:23 am Post subject: |
|
|
NP, i'm game for anything. I looked in my PM before looking at this so I owe you more info. I think I see what your saying. I'll create a user for you on my end and tell you the settings for my hotspot stuff.
BRB, getting updates from Open-Mesh now on fresh flashed OMP1. |
|
| Back to top |
|
 |
foxtroop11 Moderator

Joined: 22 Mar 2009 Posts: 763 Location: Ansbach, Germany and sometimes the States
|
Posted: Fri May 22, 2009 9:13 am Post subject: |
|
|
Ok, think I see some of the problem.
Here is a portion of the recieved.clean from open-mesh
| Code: | #@#config cp_switch
main.which_handler 5
#@#config chilli
agent.service otherchilli
agent.radiusserver1
agent.radiusserver2
agent.uamserver
agent.uamurl /hotspotlogin.php
agent.uamsecret
agent.radiussecret
agent.radiusnasid
agent.uamurlextras
agent.uamdomain
agent.custom1
agent.custom2
agent.custom3 |
now is here almost the same portion from meshconnect, I notice this main.which_handler is 6 and not 5. Wonder if that is part of the problem.
| Code: | #@#config cp_switch
main.which_handler 6
#@#config chilli
agent.radiusserver1
agent.radiusserver2
agent.uamserver
agent.uamurl /hotspotlogin.php
agent.uamsecret
agent.radiussecret
agent.radiusnasid
agent.uamurlextras
agent.uamdomain
agent.custom1
agent.custom2
agent.custom3
agent.service otherchilli
#bogus2 772827811 |
|
|
| Back to top |
|
 |
funchip Moderator

Joined: 12 Sep 2008 Posts: 253 Location: Sacramento, California, USA
|
Posted: Fri May 22, 2009 6:31 pm Post subject: |
|
|
I think @foxtroop11 has nailed down to a key point. Currently, MeshConnect dashboard sets the "cp_switch -> which_handler" field by
5 for CoovaAAA
6 for OtherChill
That could be incorrect. Should the field be set to "5" in both cases, or the value needs to be flipped...
Antonio, can you help?
Thanks! |
|
| Back to top |
|
 |
foxtroop11 Moderator

Joined: 22 Mar 2009 Posts: 763 Location: Ansbach, Germany and sometimes the States
|
Posted: Fri May 22, 2009 7:20 pm Post subject: |
|
|
| Dont think that's it. I just did further testing and although your dash sends out an alternate number setup it's actually doing the same thing as Open-Mesh. It's something else as I went in and flipped the number and it then made the node use coovaa aaa which actually worked but still did the 10 minute thing and seems to be coova restarting. |
|
| Back to top |
|
 |
|