@Gertjan:

A DSL PCI card modem device ?!! Why not, as long as it exposes on the PCI side a compatible Ethernet interface.

Well, if I remember this one right, they were offering special drivers for some brand(s?) that had to match built numbers of the distribution etc.
I doubt they were serving an ethernet-like interface to the host machine.
And manufacturers were quite limited. Very limited at that time (I haven't looked since then).

jahonix

FTP Sesame handles bridged interfaces.  This should be working.

Thankyou!

Sorry,
I have just tried what you suggested and still does not work.  Due to this not being connected to the internet what should be used as the gateway?  i have tried setting it to 0.0.0.0 and the WAN ip address but no luck either.

Thanks

Beau

@sullrich:

Take a look at /etc/inc/pfsense-utils.inc

In paticular:

enable_rrd_graphing()

Thank you sullrich for the tip.

I am using pfsense in a very pequliar way, you can find in :
http://forum.pfsense.org/index.php/topic,772.msg4635.html#msg4635

Since my original post, the ethernet card of the LAN interface, totally useless to my wireless community application, gave it's seat on the PCI bus to a 3rd wireless card, shrinking to a vlan opt that serves nothing.

This, as an update.

The important thing is that by looking at your mentioned file, I now have quality rrd graphing on one of my wireless links.

My coding skills are just not existing, so my implementation is the "crude" way.

a) For my setup the quality of the WAN interface is of minimal importance. It is directly wired to the cisco router.

So, I changed the if statement :

if("$ifname" == "wan") {
/* create link quality database */

to the Atheros card interface :

if("$ifname" == "opt2")

b) Then, I had to change the ip address it is pinging to. I did this by "hardwiring" the ip of my peer's wireless card.

I know, it's not a very educated way, but it works.

out=$btick ping -c $numpings -q $gatewayip $btick

this was changed to :

out=$btick ping -c $numpings -q aaa.bbb.ccc.ddd $btick

The next thing I have to figure out is which part of the code I have to duplicate to provide the same service for the other two wireless interfaces.

Any help on this will be appreciated.

If I had the required programming skills, I would have included a parameter box in the wireless interface web page for setting the ip to be pinged.

But I am not a programmer, nor a purist, so even so it's a nice thing to have.

It turns out to be a bad setting in the slow machines BIOS….

There is no problem with pfsense, I have problem with my hard drive. Thx for help.

Can I get access to this machine?

Great!  Good to hear.

Thanks, I've add this:

http://faq.pfsense.org/index.php?sid=147209&lang=en&action=artikel&cat=1&id=178&artlang=en

"and Luiz should be installing the .po files soon."

Still have not heard from him…

Oops, missed your reply before my edit…

Anyway, reapplied 09-06-06 and then followed your directions and voila!! It works!! Awesome!! You rock!! :D

Thanks!! 8)

this is a windows issue i have lots of experiance  with windows clients and pppoe so you have to believe this some clients will work fine and some don't
you really need to use a good pppoe dialer on any version of windows.
dr tcp will help you lower the mtu so you client does not have so many problems
i would also suggest you disable your ccp extension son the pppoe server on pfsense. i can't remember the config paramators for mpd daemon but the docs for it are on sourceforge

also another issue we see a lot is the defaults for windows pppoe auth are wrong for our setup and need to be manually adjusted.

my two cents worth i have a bit of development time at the moment so should be around on the forums for the next few days if you have other questions

It changes the IPs inside the ftp protocol to make it work with NAT and punches dynamically ports open that are needed for the ftp transfer. ftp doesn't happen on one port only (TCP 21 is only the controllsession).

Yes, you only can shape between 2 interfaces. However the problem with your setup is the following:
If you shape at the LAN side of the gatewaybox you can shape the overall bandwidth of all WANs of the gatewaybox only. So let's say each WAN has 1 mbit/s upd and down to keep it simple for calulation. This means your overall upstream isw 4 mbit/s. Now your shaper let's a single connection go out with 4 mibt/s. Now that single connection only can use 1 WAN at the same time, so it will max out the line at 1 mbit/s though it is allowed to use 4 mibt/s at the shaper box. This will overload 1 line whereas the other 3 lines are still idle. It won't work efficiently in that scenario. To make this work with multiple boxes you would need one gateway and 4 shapers at eah WAN of the gateway. This is an ugly setup and I agree to that, however it's the onyl way to do this right with multiple boxes. I have played around with custom shaperrules and a 2 WAN, 1 LAN setup but haven't managed to get it working the way I wanted it to work. However there are people reporting some kind of success with custom rules and multiple WANs. We'll hopefully have a multi interface shaper after 1.0 is out but trafficshaping gets pretty complex when using multiple interfaces so there is no timeframe for that feature yet.