Kubernetes 专家 — 诊断故障，守护集群
在真实集群上处理 NotReady、etcd 恢复、驱逐、排空与升级
고급 · 课时 15 · 实验 5
课程大纲
节点处于 NotReady
控制平面与 etcd
资源、QoS、OOM 与驱逐
调度 — Pending、排空与优先级
零停机小版本升级
参考文档
- Kubernetes — Version Skew Policy
- Kubernetes — Upgrading kubeadm clusters
- Kubernetes — Operating etcd clusters for Kubernetes
- Kubernetes — Certificate Management with kubeadm
- Kubernetes — Node-pressure Eviction
- Kubernetes — Pod Quality of Service Classes
- Kubernetes — Resize CPU and Memory Resources assigned to Containers
- Kubernetes — Pod Topology Spread Constraints
- Kubernetes — Pod Priority and Preemption
- Kubernetes — Specifying a Disruption Budget
- Kubernetes — Troubleshooting Clusters
- etcd — Maintenance (space quota, defragmentation)