Kubespray etcd nodes management
Opened this issue · 0 comments
haminhcong commented
Etcd certs
Replace unhealthy etcd node
Two methods
- Recover control plane: https://github.com/kubernetes-sigs/kubespray/blob/release-2.15/docs/recover-control-plane.md
- Remove etcd node then add etcd node: https://github.com/kubernetes-sigs/kubespray/blob/release-2.15/docs/nodes.md
Backup etcd data & certs
- https://codefarm.me/2019/05/22/kubernetes-recovery-master-failure/
- https://medium.com/velotio-perspectives/the-ultimate-guide-to-disaster-recovery-for-your-kubernetes-clusters-94143fcc8c1e
Disaster recovery scenario
- Write your own
- Based on playbook
https://github.com/kubernetes-sigs/kubespray/blob/release-2.15/roles/recover_control_plane/etcd/tasks/recover_lost_quorum.yml