(moderator: the good bits at the end)

I've been struggling with my 2820vn since I've owned it. It hasn't really worked in since I've owned it which must be more than 12 months. Constant re-connects, reboots and lock ups.

There are loads of possible reasons for why. I live in an area drowning 2.4GHz APs. I share my bandwidth with nearly every student at De Montfort University on the Montfort exchange. That exchange has just gone 21C and maybe suffering for it or the dodgy Cisco routers they might be using.

When 3.3.3 came out I thought I may as well give that a go. No success.

So, I thought, sod it I'll put a 54g PCI card, I have knocking about, into a Linux box and configure it as an AP and then plug in the cheap modem my provider gave me and see what's what.

I had to, turns out I didn't have to but I did, update my Fedora install to get least effort latest kernel installed for the lastest wifi AP supporting drivers working. I then configured hostapd, a bridge and, hey presto, I had a working wifi setup.

The next step was to look at getting the cheap modem working. As I started this step I checked one more time the status page for the 2820vn. It had been going 33 hours without a break. A record for this device. It's now 55 hours and counting. And I have a theory for this. From the change history for the firmware you get the idea that the software is a little shaky on volume connections like nmap port scanning. I had a squid web proxy running, on the server I upgraded, which refused connections at the server and not the 2820vn. I am coming to the conclusion that the volume of traffic trying to use the proxy is the reason the 2820vn has been crashing.

I guess I can test the theory but for now I am happy that the 2820vn hasn't crashed for 2 days.