2023 Router tweaks

So time for the quarterly set of tweaks to see if I can improve my speeds, reliability etc. This time I’m trying to sort out an issue with Wi-Fi which is causing a server, located in my garage, to be unresponsive on the network. The first set of changes were nothing to do with the router but to qBittorent, in my aim to maximise the upload speed I’d allowed the software to swamp the local network so the Server wasn’t actually able to connect to the network and get an IP address! Some of these tweaks are unique to the Billion BiPAC 8900AX I use while others are standard and so relate to all, I’ll try and identify which is which.

MMM  DD HH:MM:SS kern warn kernel: Missed Membership report from for 1 times. Waiting..

So in the router logs I frequently saw the above. The 169.254.149.xxx means the Server hasn’t actually been assigned an IP address yet, have a search for that range to find out more details. Once I’d reduced the number of connections allowed for all torrents and active seeding torrents I started to see less of those issues, they still occur but at least the Server does get assigned a local IP address during the day now. For the Billion router you can QoS rules, under Configuration->Quality of Service. For these to work for torrents I had to pre-configure the port to use on inbound and outbound connections, so I don’t use UPNP, which isn’t a bad thing it seems, again search for security issues round that. So my torrent connections are now set to a low priority on the outbound (upload/LAN to WAN) side and limited on the inbound (download/WAN to LAN) side to 67% of the available bandwidth, these rules also only to the torrent server, which has a fixed IP address. This allows Streaming and browser to continue quite happily with no interruptions whereas before this stuttering was common.

Now onto the generic Wi-Fi changes…

Bizzarely this involved reducing the bandwidth settings in the 2.4GHz/5GHz ranges. For the 2.4 I now have a bandwidth of 20MHz while the 5 is currently set to 40MHz. The result is that the 2.4 has greater stability, as does the 5, but the range of the 5 has actually increased into more rooms in the house! This seems to be down to collisions with neighbouring networks, not many four at the worse case, but by trying to use the full bandwidth it was frequently causing issues. Now I’m using less of the bandwidth the quality of the connection, and range, has improved. Seems strange? Yes but I did find several sites that explain this in detail and why it can improve the quality of the connection.

I did also enable “MLD Multicast Proxy”, this is the IP6 version of “IGMP Multicast Proxy” it turns out, I wasn’t aware until I started searching for improvements. But has it made a difference? I can’t tell but it hasn’t made it worse and I do know that enabling IGMP solved our initial issues, when I first started using the Billion, when trying to use Youview and the online TV catchup services.

Lastly I reduced the “RTS Threshold” from the maximum 2347 down to 2040. Again no noticeable improvement but it’s not worse. This is my current focus though, so does it need to go lower? I’ve seen recommendations to drop it to 500 but these are accompanied with the caution to do so gradually as that doesn’t suit all. So, hopefully, I’ll get some updates to this later.


A tale of two drives

So just happened to spot, when running HWiNFO64 to check motherboard version, that my 250GB Crucial MX500 is down to 8% life remaining! What? Couple of months of go when I was checking that value it had 28% life remaining, so all good and plenty of time to plan my next upgrade. Now I’m in panic mode and ordering a new 500GB Crucial MX500 for £43, which is a good price as the 250GB cost me £63 just over four years ago.

HWiNFO64 SMART status
SMART Status from HWiNFO64 v7.26

So what happened? Seems that the Crucial drives have a potential with many small packet writes, this may just be on the model made all those years ago and not on the current models, Crucial have always had a good reputation from reviews. So in the SMART details we can see it has written a lot of data, compared to the other MX500. The failing drive is my boot disk so has the pagefile, core programs (Anit-virus, VirtualBox, Browsers) and applications (Office, VLC, Notepad++) on it.while the other SSD, only a couple of months newer and running the same firmware version, has all the games, documents, photographs and BOINC application.

So the failure drive presumably has all the small packet writes (temporary files, browser cache & cookies, pagefile) while the still okay SSD has the larger writes (photographs, documents).

So just about to install the new 500GB Crucial MX500 replacement drive, and we’ll see how that goes over the coming month. After that the PC will probably be passed on as I should have my new one built and moved into the world of NVMe drives.

Quick update to this. The machine, Treacle, has now been retired and relegated to a hardware test role, admittedly old hardware. The new Crucial MX500 drive is at 100% life remaining while the old 1TB MX500 has only 64% life remaining, so still pretty good.

So is it aceptable to sell an SSD that has only 8% life remaining? It will need a secure erase, which presumably will reduce that further, cue another addition to this post.

The secure erase, three passes using “British HMG IS5” via Eraser an overnight job, didn’t reduce the disk life any. So I put the drive on eBay with full disclosure of the remaining drive life, including screen grabs from HWInfo SMART data, and a starting price of £3 + £3 p&p. One week later it sold for £26 inc p&p, when you can buy a brand new one for £38.08 inc p&p! The buyer, one of two bidding, was happy with the drive and condition. So why pay, what I consider, over the odds? Did they expect to be able to retrieve data from the drive? I have had a close friend, computer illiterate it now seems, sell a computer after physically destroying the drive only to be contacted by the buyer to say they have his personal data now (The buyer did send the drive back).

So is there a market in second hand drives purely for what data you can retrieve from them? I know it’s possible to get data from incorrectly erased drives, I’ve had to recover said data for friends in the past, so is that data of use to someone? You’d need some serious time and software to be able to make use of any account/password data on the drive.

Enabling WPS on a Billion BiPAC 8900AX

By default WPS is disabled on the 8900AX, so if you need to connect a device, TV, printer etc, using WPS then you will have to go into the Configuration menu and enable it.

  • This can be found under the Configuration->Wireless 2.4G->Security menu.
  • Select the “Enable” value for the “WPS” combo, under “WPS Setup”.
  • Ensure that the “Add Client” option is set to “Use STA PIN”.
  • Make sure that the “PIN” value is blank.
  • Leave all other options at their default settings.

Once this is done all you need to do, to connect the device using WPS, is to press the WPS button on the router, for around a second, and then perform the equivalent activity on the device you want connected.

Windows 10 restarts after shutdown

Currently have an issue where most of the time shutting down the PC just performs a restart, annoying to say the least, been going on since late 2018.
I’m currently trying out this solution to see if it can be resolved.

6-Dec-19: A couple of test shutdowns later and it seems to be working, one to monitor.

Windows PC not auto-booting?

Since I installed the “Windows 10 Fall Creators Update”, build 16299, the desktop has stopped auto-starting, to it’s BIOS configured schedule. It’s scheduled for a 5:01am start, but every morning I’ve come into the room, since the update, it’s sitting there switched off.

A couple of Google’s later and here’s the answer, well I hope it is as I haven’t verified it yet, something for tomorrow morning. So the suggestion is to disable Windows “fast startup” mode, details on doing that can be found over at TechAdvisor.co.uk, and a more detailed explaination of why is at HowToGeek.com.

29th October 2017: Seems to have resolved the issue.

Cannot synchronise MLAT with Piaware?

Possibly an issue with Pi firmware of 4.9.x, find out your version by using this command

uname -a

If so then issue these commands, the first installs the firmware updater the second the 4.4.50 version of the firmware.

sudo apt-get install rpi-update
sudo rpi-update 52241088c1da59a359110d39c1875cda56496764

If you can’t install the firmware because you have a “custom” kernel then you’ll have to get the latest version of Raspbian and install that, then run the above two commands.

Reducing wear on your sd-card

One way to increase the lifespan, and presumably, reliability of the sd-card, currently being hammered, in your Pi is to put some of the temporary files into memory.

Add the following to your /etc/fstab

tmpfs /var/log tmpfs nodev,nosuid,size=50M 0 0

If the size isn’t specified then half of the machine’s memory is used.
nodev and nosuid are security options, the suggestion is that these only need setting for public facing machines.


8-Jan-2018: If you’re using the lighttpd web server then be aware that it stores files in /var/log, and it doesn’t like them being removed. This will cause the server to fail to start.