Hello all,

I have an x86 machine setup as a bandwidth shaper/firewall.
The machine has 2 Intel 10/100/1000 NICs.
The ports are bridged along with some VLANs.
We then use Simple Queues with bursting for traffic management.
We also use firewall rules with access lists to allow active customers through and block everything else.

We decided to use ROS6 for it's supposed better simple queue performance.

Every release we have used, including RC11 will randomly stop passing all the bridged traffic.
I'm not sure if the problem is the bridging, VLANs, SImple Queues, or firewall rules.
It could work just fine for anywhere between 10-30 days or so.
Then all devices behind the box are unreachable.
You can still log into the bridged management interface and reboot it.
After the reboot everything returns to normal.
I have not had a lot of time to analyze things when it breaks because it takes down all of our customers behind it.

I just wanted to post this because I know there have been similar reported problems with earlier RCs.
It is still happening in RC11.

Did you ever figure out the cause? I have a wireless link which does this. Problem disappeared for a year or better then with V6.x it seems to be back. I am running a RB433AUH with two wireless cards. This gives me over 100 meg accross the wireless interfaces. I can always reach the tower from the NOC, but not beyond it. Looking in the registration table, everything is still connected, just not passing traffic. To fix it, i either reboot the tower or simpy remove the link in the registration table. As soon as it reconnects and the the WDS is reistablished, traffic flows again. And we seem to have this happen, when it happens every 24 hours.