A homeowner ran a single bare-metal Debian server for 145 days (about 56 TiB of traffic) that quietly hosted NAS, NVR/video analysis, TV streaming, home automation, DNS, MariaDB, Prometheus/Grafana, UniFi controller, development environments and a collection of Docker containers where appropriate. The upgrade from Debian 12 (Bookworm) to 13 (Trixie) serves as the central claim: modern Linux reliably consolidates many workloads directly on hardware without reflexively layering hypervisors or containers. Managing the machine with Ansible, infrastructure-as-code, monitored services and deliberate repositories prevents it becoming a “snowflake” and changes failure recovery from repairing an idiosyncratic install to reproducing a known state.
The upgrade was done carefully: repositories consolidated to deb822/deb.debian.org, backports and unused sources removed, staged apt operations performed and reboot delayed until recovery was assured. Real problems that surfaced were concrete and fixable: ZFS DKMS failed to build against the new kernel and was resolved by installing updated zfs packages; NodeSource repo signatures were rejected for SHA-1 and fixed by replacing the key and moving to Node.js 24/Yarn 4; Java runtimes were consolidated to OpenJDK 21; pipx environments rebuilt for Python 3.13; Ansible modules modernized. After reboot the server ran with much lower load, about half the memory use and >10 GB of obsolete packages removed, demonstrating upgrades both expose configuration debt and create opportunities to eliminate vendor-specific cruft.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.