Page 3 - Linux

AI generated

Exporting Bash Script Metrics to Prometheus: The Node Exporter Textfile Collector

Maxim Mironjuk - April 08, 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.

Linux

AI generated

Understanding OverlayFS: The Union Filesystem Behind Every Docker Container

Maxim Mironjuk - April 06, 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.

Linux

AI generated

systemd-nspawn: Lightweight Containers Without Docker

Maxim Mironjuk - April 03, 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.

Linux

AI generated

Understanding the Linux Filesystem Hierarchy (FHS)

Maxim Mironjuk - March 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.

Linux

AI generated

systemd vs. OpenRC vs. runit: Comparing Init Systems for Servers and Containers

Maxim Mironjuk - March 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.

Web Performance Linux

AI generated

CPU Governor and Frequency Scaling Tuning for Latency-Sensitive Servers

Maxim Mironjuk - March 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.

Linux

AI generated

Analyzing Kernel Panics and Crash Dumps with kdump

Maxim Mironjuk - March 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.

Linux

AI generated

Debugging Network Connections Step by Step

Maxim Mironjuk - March 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.

Bash Linux

AI generated

Sort, uniq and comm for Comparing Datasets

Maxim Mironjuk - March 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.

Bash Linux

AI generated

wait -n for Parallel Job Harvesting in Bash

Maxim Mironjuk - March 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.