ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
DevOps & observability · #k3s
When a K3s cluster loses API access after a node restart, the embedded SQLite etcd datastore is often the hidden cause.
Choose RKE2 for general production clusters and K3s for edge or memory-constrained nodes; never mix distributions in one cluster.
Decide when to keep K3s embedded etcd and when to move to a dedicated external etcd cluster for control-plane high availability.
Run K3s on immutable k3OS nodes by configuring through cloud-init, protecting the cluster token, and verifying writable state partitions.
Learn how to boot k3OS from a USB stick with a persistent data volume, configure a lightweight k3s cluster on embedded hardware, and understand the trade‑offs and practical steps for reliable edge deployments.
Architecture note on K3s embedded SQLite for single-node clusters: requirements, minimal design, trust boundaries, operational checks, failure modes, and migration triggers with qualified statements and uncertainty markers.
K3s replaces etcd with SQLite by default, saving memory and operational overhead on edge devices. This post explains why that works, when it breaks, and how to migrate to PostgreSQL for HA — with a concrete step-by-step example.
A diagnostic guide for resolving K3s HA control-plane join failures, etcd quorum issues, and raft membership errors in versions v1.28–v1.30.
Deploy a minimal, highly available k3OS cluster using its built‑in etcd. Learn the exact install flags, trust boundaries, operational checks, and failure‑mode mitigations for a 3‑node HA design.
kubernetes
rancher
k3os
sqlite