Managing cloud-native distributed block storage for Kubernetes with Longhorn.
homelab-journey
Replacing misleading volume-level Longhorn alerts with disk-level rules, recalibrating snapshot overhead to 50%, and how an unrelated incident accidentally produced the baseline data needed.
homelab-journey
Empty Longhorn panels led to metrics-server, a Talos cert quirk, a reboot that broke Garage, and 16 silent hours of failed backups.
homelab-journey
prometheus-server at 90% allocated, 5.6 GiB of real data. The alerts fired — but were they firing on the right thing? Snapshots, unreclaimed blocks, trim, and a PromQL join.
homelab-journey
Longhorn allocation never coming down? Here's how to enable automated fstrim via a RecurringJob on Talos Linux — and what it can't fix.
homelab-journey
Rolling NVMe upgrade on a live Talos cluster: phase=failed, the extraMounts unlock sequence, and why storageReserved needs an explicit value.
homelab-journey
Three new RK1 worker nodes, three different Longhorn disk problems. The right way to add a node — and what happened when I didn't follow it.
homelab-journey
Upgrading Talos, Kubernetes, and thirteen Helm charts with real workloads running — the safe order, the tools, and what went wrong. The how, not the what.
homelab-journey
The controller managed its own backups perfectly. It had no idea about the rest. v0.11.0 adds an audit script and two new alert conditions to close that gap.
homelab-tools
stern tails logs from multiple Kubernetes pods at once. I used it to find a pending Traefik upgrade and reconstruct a PVC resize across five pod types.
homelab-journey
New PVCs were silently unprotected until I noticed. I built a label-driven controller to close the loop — jobs created automatically, violations alerted, archives handled.
homelab-tools
k9s gives you an interactive terminal UI for your Kubernetes cluster. I installed it and immediately found a backup job that had been silently failing for eleven days.
homelab-journey
Importing Longhorn, Kubernetes, and resource dashboards onto the Bletchley cluster — and fixing the scrape config that was only collecting from one of four nodes.