If you're using Docker healthchecks with an autoheal sidecar (like willfarrell/autoheal) to auto-restart unhealthy containers, there's a class of failures they'll never catch.
A common pattern for self-healing containers:
services:If you're using Docker healthchecks with an autoheal sidecar (like willfarrell/autoheal) to auto-restart unhealthy containers, there's a class of failures they'll never catch.
A common pattern for self-healing containers:
services:Hetzner's cheapest VPS plans come with IPv6-only connectivity. Docker doesn't handle this well out of the box — containers won't be able to reach the internet. Here's the full setup.
Docker defaults to IPv4 networking. On an IPv6-only host, containers can't resolve DNS or reach external services (package registries, APIs, OAuth providers, etc.). You'll see:
apt-get update failing inside containersIf you're running Tailscale with an exit node (e.g. Mullvad) on a server that also runs Docker containers, you may hit a maddening issue: containers appear healthy but are unreachable from the host.
curl http://127.0.0.1:<port> to a Docker container hangs (TCP connects, but HTTP times out)docker exec into the container and curling localhost works fineThe default kernel OOM killer reacts too late for desktop use — by the time it fires, the system is deep in swap thrash and unresponsive. systemd-oomd is better but still too conservative for desktops with lots of browser tabs.
earlyoom monitors memory and swap usage and kills processes before the system becomes unusable.
# Fedora / rpm-ostreeApps (especially Flatpaks) take ~2 seconds to launch because the NVIDIA dGPU enters D3 power state when idle. Waking from D3 adds ~2 seconds of latency to any GPU operation.
This affects hybrid laptops (NVIDIA + AMD/Intel iGPU) where the kernel's runtime power management suspends the dGPU.
The official Dropbox CLI (dropbox.py) doesn't detect the Flatpak Dropbox daemon. Commands like dropbox status report "Dropbox isn't running" even when it is.
Root cause: The CLI checks ~/.dropbox/dropbox.pid, but the Flatpak daemon writes its sandboxed PID there, which doesn't correspond to a real process on the host.
Tested with Goodix 27c6:550a on Fedora 43 (Bluefin DX, nvidia-open). Should work for other Goodix TOD-compatible readers.
The Goodix driver requires libfprint-tod, which replaces the base libfprint package.
# Add the COPR repo (uses $releasever automatically)TL;DR: Qt6 Flatpak apps using the KDE runtime ignore GNOME's dark mode preference. The workaround is three environment variables + Kvantum.
If you run Qt/KDE Flatpak apps (qBittorrent, OBS Studio, VLC, etc.) on GNOME with dark mode enabled, they render with a light theme — white backgrounds, light widgets, the works.
This is caused by two bugs in the Qt/KDE stack.
If you're using Chrome installed via Flatpak (common on Fedora Silverblue, immutable distros, etc.), the Claude Code browser automation won't work out of the box. Here's how to fix it.
Claude Code installs the native messaging host to the standard Chrome location, but Flatpak Chrome:
/tmp that isolates the Unix socket used for communication~/.claude/ or ~/.local/share/claude/ by default