Hello

I have a Verizon FIOS router A90-9100EM15-10 made by Westell. It runs on the Linux Kernel. It's worked fine for me for quite some time until recently. I have found several odd quirks that have resulted in me getting banned from one particular website/forum. I talked to the web admin and he said that he has a script that auto bans if too many connections are opened. He doesn't mind un-banning me, but this has got to be aggravating for him, it certainly is for me. He's tried white listing my IP but for whatever reason that doesn't work either.

I have opened a few tabs to various threads in the forums that I am watching and from time to time I reload the threads to see if there are updates. It's at the reload that the issue seems to happen.

I looked at my router's firewall and it says I have hundreds of connections to his server, but on my local machine it only shows 6 -8 connections. Also through the router's firewall it says that the "Time To Live (seconds)" is almost 432000 seconds, that is just shy of 5 days. This seems to be the case with almost every website I visit, but he is the only one that bans, and I spend more time there than others. I have tried this from my LinuxMint laptop, my Scientific Linux box and also from our Mac, they all have the same outcome, so it definitely seems like a router issue for me. Also some are plugged in and some are wireless so that takes another variable out of the equation.

My thought is that instead of re-using connections it's just opening a new connection and since the timeout is so huge they aren't timing out quick enough.

I have tried defaulting the router to factory settings, upgrading firmware, downgrading firmware. Nothing makes a lick of difference. I went to Westell's website and found that they have the source code for download. http://westell.com/support/ultraline...90-9100em15-10 So I thought I'd search the code to see if there was anything that jumped out at me. I actually found a few things that I wanted to ask. I've only been into Linux for the past year so I'm pretty newb to some of this.

From what I can see the firmware seems to be built with the Linux kernel version 2.4

TTL is usually a max of 255, but why is this labeled as seconds? TTL as far as I know it is just a hop count. I thought about changing this to some lower value to see if it changes the operation on the router, but since 255 is the standard I'm leery of that.

This was interesting, I have no idea what it means, but 600 seems to be the default in a lot of documents I have found on the web and on my computer, if it is multiplied by whatever HZ is this can grow quite quickly... I couldn't find any reference to what HZ actually was, maybe this was defined in one of the includes in the header, I'm not sure. I thought about setting this to the default of 600 that my LinuxMint computer is set to. Any comments? Perhaps this code works fine on a real computer but this router has an Infineon twinpass processor and it isn't quite standard, that might be why it doesn't work right

If this is a real firmware issue I wonder why nobody else has noticed it, or am I just on the wrong track here? I'm tired of getting banned! My last ban had over 500 connections



Thanks!

I've been reading a bit more on the web and reading from one forum to the next I found a link back to here

http://www.linuxquestions.org/questi...e-full-581142/

In this post someone mentioned a current setting of "net.ipv4.netfilter.ip_conntrack_tcp_timeout_established = 432000" Which is exactly the number that I said my router is reporting. So it seems that keeping established connections open for 5 days is some sort of standard?

https://forum.openwrt.org/viewtopic....d=29385#p29385

This post from an openwrt developer says that "By default, OpenWrt (as of White Russian RC5 and latest Kamikaze build) defaults to tracking established TCP/IP idle connections for 5 days. This is the standard in the *nix world."

Why in the world keeping the connections open for 5 days is a "standard" is something I do not understand.

So perhaps part of my "Problem" wasn't actually a problem. But the crux of the problem is that it doesn't look to be reusing connections.

Maybe this comes to standard web browser architecture? Does anyone know how a browser reloads a page Vs clicking a link? If I click a link does it reuse a connection and reload open new connections?