Proxmox

Every DNS lookup in my k3s cluster was appending 'home.lan' and recursing externally. I blamed DHCP, wrote a Terraform fix, and watched the plan show zero diff. The real source was three layers up.

When Ollama, Minecraft, and etcd compete for CPU on the same host, the scheduler doesn't know which one is critical. Proxmox cpu.units gives etcd 2x priority — and without it, leader-election timeouts cascade into cluster instability.

Jellyfin runs in a GPU-passthrough LXC for hardware transcoding. SABnzbd, Sonarr, Radarr, and Bazarr run on a separate LXC with per-flow traffic isolation. A Python CronJob in k3s ties them together. Here's why the media stack left Kubernetes.

All VMs and LXCs starting simultaneously spiked the Proxmox host load to 147. The fix was NFS first, k3s nodes 30 seconds apart, and LXCs last — a boot order strategy that costs nothing but prevents boot storms.

VM dedicated memory is a real reservation. LXC dedicated is a soft ceiling. Summing them together overstated pressure by 40GB — but the real risk was underestimating it. Here's the two-tier memory model that prevents a repeat of the ZFS ARC freeze.

My Proxmox host froze under load three times in one week. The root cause wasn't software — it was a marginal PSU triggering TDP throttling, which exposed a ZFS ARC deadlock. Here's the full chain from symptom to fix.

DISASTER-RECOVERY.md documented the Proxmox Backup Server restore procedure in detail. It had never been executed end-to-end. Running it for the first time - against a scratch VM, alongside the still-running original - found a network conflict the documentation never mentioned.