Linux Security

Why Move /var/log to a tmpfs?

On most production boxes and home‑lab rigs the root file system is a single ext4 or btrfs partition that also holds /var/log. After a few months the logs can eat up gigabytes of space, especially on servers that churn out audit, kernel, or app logs. A full root partition can block package upgrades, kernel updates, or even prevent a boot if the filesystem gets remounted read‑only. Moving /var/log to a tmpfs frees the disk right away, keeps the logs in RAM, and cuts out the need to compress or rotate them manually.

Why a kernel update can drop you into emergency mode because /boot won’t mount

A fresh kernel can change the initramfs or the boot‑loader’s config, and if the loader can’t find the /boot partition the system drops into that classic emergency shell:

Emergency shell (root=rw) – press Ctrl‑D to continue

Below is a no‑fluff, hands‑on walk‑through that keeps the commands you need, the trade‑offs, and a few security reminders.

1. Inspect the emergency environment

First, make sure the culprit is really the /boot mount. In the emergency shell run:

I’m ready to help refine the article, but I’ll need the original text first. Could you paste it here?

I’m ready to help polish it, but I’ll need the article text first. Could you paste the draft you’d like me to rewrite?

When a VPS loses its /etc directory, the system can become unbootable or at least unusable until the configuration files are restored. If you have been running Borg backups (the recommended tool for incremental, deduplicated snapshots) the recovery path is clear: locate the snapshot that contains the last good /etc, extract it, and replace the missing directory. This walkthrough shows the exact commands, trade‑offs, and security checks you should perform on a current 2026 VPS environment.

Why Bluetooth headphones keep dropping on GNOME 48

Bluetooth audio on GNOME 48 is now handled by PipeWire instead of the old PulseAudio stack. The new stack is cleaner, but it also turns on a “low‑power” mode that can make a headset drop after a few minutes of idle time or when you close the laptop lid. The real culprit is usually the kernel’s power‑management for the Bluetooth controller and PipeWire’s default settings.

The underlying problem

When a systemd timer stops firing after a kernel upgrade, the first thing that comes to mind is “something broke during the reboot.” In reality the timer is still in the unit files; it’s the time‑keeping that gets reset. The kernel’s real‑time clock (RTC) is re‑synchronised, and any timers that were scheduled for a time in the past are dropped unless you tell systemd to treat them as persistent.

Quick Memory Hunt with systemd‑cgtop

When a server starts to feel sluggish, most folks jump straight to CPU or disk I/O. Memory is the silent saboteur that slips under the radar until it’s too late. On modern Linux, every service lives in a cgroup, and systemd‑cgtop gives you a live, per‑service view of how that cgroup is eating RAM. Unlike top or ps, it aggregates the numbers for you, so you can spot the offender in a glance.

Why Pinning the NVIDIA Driver Matters on Ubuntu 24.04

Kernel updates are a blessing and a curse. They bring new features, hardware support, and security patches, but they also can break proprietary modules that depend on a specific ABI. The NVIDIA driver is a prime example: a new kernel may require a new driver build, and if you’re running a GPU‑heavy workload or a GPU‑accelerated service, an unexpected reboot or a “module not found” error can be costly.

The first time I noticed the problem, a Docker‑based microservice was silently timing out on every outbound request. The logs were full of “no servers could be reached” errors, but the host machine could resolve the same hostnames without issue. After a quick check of the container’s /etc/resolv.conf, I saw an odd line: search mydomain.com. The host’s /etc/resolv.conf had no such entry. That single missing search domain was the root of the DNS failure inside containers.