restic: Backups That Deduplicate, Encrypt and Verify
One binary, an encrypted repository on almost any storage, and snapshots that only store what changed. What restic is good at, the retention policy that actually frees space, and when…
systemd Beyond Services: Units, Targets and Dependencies
systemd manages units, and services are only one kind. Sockets, mounts, targets and slices explained — plus the difference between requiring something and waiting for it, which is behind most…
How Packages and Dependencies Actually Work
What a package contains, how a repository is indexed and trusted, why shared library versions are the real reason dependencies exist, and what mixing sources does to a system.
Containers: What Is Actually Happening
A container is an ordinary Linux process with a restricted view of the system. Namespaces, cgroups and overlay filesystems explained — and the problems that stop being mysterious once you…
What the Kernel Actually Does
Linux is the kernel — not the distribution around it. What lives on each side of the userspace boundary, how to watch a program ask for something, and why an…
Shell History — Stop Retyping Commands
Ctrl+R, history expansion, the readline editing keys worth learning, and the configuration that stops your history vanishing every time you close a terminal.
User Management — useradd, usermod, passwd and chage
Creating accounts, adding someone to a group without wiping the groups they already had, password ageing with chage, and why locking an account beats deleting it.
systemd Timers: Scheduled Jobs That Tell You What Happened
Timer units do what cron does, with proper logging, dependency handling and catch-up after downtime. The syntax, a working example, and when cron is still fine.
dd: Writing Images and Cloning Disks
Write a bootable USB, clone a disk, or destroy the wrong one — dd does exactly what you tell it, without asking. How to use it, and how to check…
Monitoring a Server Without a Full Observability Stack
Find out a service died before your users do — systemd OnFailure alerts, a disk-space check, and external uptime monitoring. No Prometheus, no Grafana, about an hour.