Field notes on systems, tools, and practical experiments.
Small systems stay useful when their important parts can be explained on one page. A short inventory of services, ports, files, and recovery steps is often worth more than a complicated management layer.
The goal is not to avoid change. It is to make every change visible, reversible, and easy to understand a few months later.
A fast result from a single test says very little about a network. Stability over time, packet loss, route changes, and recovery after an interruption usually matter more than the best round-trip number.
Good measurements are repeated from the places where a service is actually used. They record failures as carefully as successes.
The most dependable tools are rarely the ones with the longest feature lists. Clear configuration, readable logs, and an ordinary backup process make a tool easier to trust.
Reducing the toolbox also reduces the number of assumptions that must be remembered during an incident.
A backup is useful only when it lives outside the system it protects and can be restored without guesswork. Configuration and recovery notes deserve the same attention as application data.
A small restore test answers more questions than a large collection of untested archives.
Routine work should be short and predictable: check health, read recent warnings, apply one deliberate change, and verify the result from outside the machine.
When the process is simple enough to repeat, maintenance stops becoming a special event.
KSO70 is a small notebook about practical infrastructure, careful experiments, and the habits that keep technical systems manageable.