Linux: Guides and Practical Know-how | Page 3 | mironsoft Blog
Linux
-
Apr 8, 2026
A cron job that runs at night normally vanishes without a trace for Prometheus: no running process means no scrapable HTTP endpoint. The Node Exporter textfile collector solves that problem by reading simple .prom files from the filesystem that a Bash script writes itself, making cron job duration, success, or failure queryable in Prometheus without any extra service.
-
Apr 6, 2026
Anyone starting Docker containers every day uses OverlayFS without usually knowing it. Understanding how lowerdir, upperdir, and workdir interact, and when the expensive copy up mechanism kicks in, makes it possible to design image layers more deliberately and avoid performance problems with large files inside containers.
-
Apr 3, 2026
Where Docker ships a full runtime complete with its own image format, registry integration, and orchestration, systemd-nspawn takes the opposite approach: a minimalist tool, deeply integrated into systemd, that turns a chroot-like directory or image into a fully isolated container within seconds, with no daemon, registry, or compose file involved.
-
Mar 27, 2026
Mixing up /etc, /var, /usr and /opt leads to servers that cannot be automated, packaged cleanly, or maintained predictably. The Filesystem Hierarchy Standard defines what every top-level directory on Linux is meant for, and following it gets you systems that work smoothly with package managers, systemd and automation tooling instead of fighting them.
-
Mar 26, 2026
PID 1 decides far more than just the boot sequence: service supervision, dependency resolution, and logging architecture all hang directly off the chosen init system. systemd dominates on classic servers, while OpenRC and runit hold their ground in lean container base images like Alpine and Void Linux, for solid technical reasons.
-
Mar 25, 2026
The wrong CPU governor forces a server to ramp up its clock speed on every incoming request before it can actually respond. This delay is barely visible in monitoring, but shows up in practice as fluctuating response times, especially under bursty PHP-FPM traffic.
-
Mar 23, 2026
When a server simply reboots after a kernel panic, the actual cause stays in the dark until the same crash happens again. kdump automatically produces a complete memory dump when a kernel panic hits, and the crash utility turns that vmcore into something readable, with a backtrace, process state and the exact spot in the kernel code where it broke.
-
Mar 20, 2026
A server suddenly can no longer reach a payment provider, an external API, or an internal service, and the error message Connection timed out says nothing about the cause. This article shows the complete diagnostic path for a failing network connection: DNS, routing, firewall, TCP handshake and application layer, each with the appropriate command for that layer.
-
Mar 14, 2026
Comparing two CSV exports, finding out what got added, what is missing and what is shared: sort, uniq and comm solve this task without a database and without an external scripting language, simply through the right combination of three classic Unix tools.
-
Mar 13, 2026
Classic wait blocks until every started background job is done, which caps parallelism at the boundaries of whole batches. wait -n instead waits for the first job to finish, whichever one it is, finally making worker pools with consistently high utilization practical, without reaching for external tools like GNU Parallel.