Meeting Notes Nov 9, 2021

[be170f...@gmail.com]

The meeting started promptly at 6:40pm in regular QCLUG fashion

The first topic was System Cloning with Code by Bob.

The talk was centered on a tool that he created for OpenSuSE/Zypper called Package Re-installer which he wrote in C

Bob appears to be adding support for Ubuntu/Debian to his tool

He started with OpenSuSE 15.2 and upgraded to 15.3 and continued to test his tool on.

You have to provide a couple of package lists that are created using the native package manager

The tool is invoked by running
pkg_install_v005
which creates a file called install_pkgs.sh

A script file is then output which has a zypper line with 5 packages listed per line and will install them in many transactions.

Bob has a beta version in progress with interactive prompts that asks which OS you would like the generate a package installation script

Next topic was SQM update on OpenWRT by Alex

We were gifted another take of the previous SQM presentation by Alex as a primer to his follow up SQM slides, thanks for the refresher!

Basically bufferbloat is a phenomenon that happens (in layman terms) when you go from big pipe to little pipe regarding network bandwidth.

You can avoid having the ISP throttling affect your connection by using SQM to keep from hitting the ISP throttle and you'll experience a higher quality top speed without experiencing a quality loss due to bufferbloat.

We zoomed through the refresher slides if you need more detail refer to the July meeting notes on SQM: https://groups.google.com/g/qclug-list/c/8XLTSNG44M0

Alex confirmed he still uses "cake" as the queuing discipline which has been blessed by Chris as long as your cpu doesn't tank when in use. Monitor your CPU closely, or not.

Here comes the update

Nov 2021, Metronet fiber acquired!

When he tested with SQM disabled he was seeing numbers very close to the 500/500mbps that he is paying for which is :verygood:

Next he tuned SQM for for downstream and upstream and found that is downstream bufferbloat wasn't too bad and he is able to get close to 500mbps (I don't remember what the figure wast) but for upstream he ended up without about 80mbps because he was unable to relieve the bufferbloat at a higher setting.

Alex will follow up again in the near(ish) future to let us know if he can get better upstream than that on metronet without bufferbloat but either way he is happy with his new blazing fast connection.