1
# 查看控制平面组件状态
2
kubectl get componentstatus
3
kubectl get pods -n kube-system
4

5
# 查看 etcd 健康状态
6
kubectl exec -n kube-system etcd-<node> -- etcdctl endpoint health
7

8
# 查看 API Server 延迟
9
kubectl get --raw '/metrics' | grep apiserver_request_duration_seconds

1
# 节点维护操作
2
# 1. 驱逐 Pod
3
kubectl drain <node-name> --ignore-daemonsets --delete-emptydir-data
4

5
# 2. 标记不可调度
6
kubectl cordon <node-name>
7

8
# 3. 解除维护
9
kubectl uncordon <node-name>
10

11
# 查看节点资源
12
kubectl describe node <node-name> | grep -A 5 "Allocated resources"

1
# 方式一：直接备份
2
ETCDCTL_API=3 etcdctl \
3
    --endpoints=https://127.0.0.1:2379 \
4
    --cacert=/etc/kubernetes/pki/etcd/ca.crt \
5
    --cert=/etc/kubernetes/pki/etcd/server.crt \
6
    --key=/etc/kubernetes/pki/etcd/server.key \
7
    snapshot save /tmp/etcd-backup.db
8

9
# 方式二：API Server 备份
10
kubectl get all --all-namespaces -o yaml > /tmp/cluster-backup.yaml
11

12
# 定时备份脚本
13
#!/bin/bash
14
BACKUP_DIR="/var/backups/k8s"
15
DATE=$(date +%Y%m%d_%H%M%S)
16
ETCDCTL_API=3 etcdctl \
17
    --endpoints=https://127.0.0.1:2379 \
18
    --cacert=/etc/kubernetes/pki/etcd/ca.crt \
19
    --cert=/etc/kubernetes/pki/etcd/server.crt \
20
    --key=/etc/kubernetes/pki/etds/server.key \
21
    snapshot save ${BACKUP_DIR}/etcd-${DATE}.db

1
# 恢复 etcd（单节点）
2
 systemctl stop etcd
3
 rm -rf /var/lib/etcd
4
 ETCDCTL_API=3 etcdctl snapshot restore /tmp/etcd-backup.db \
5
    --initial-cluster=new-etcd-0=https://192.168.1.10:2380 \
6
    --initial-advertise-peer-urls=https://192.168.1.10:2380 \
7
    --name=new-etcd-0 \
8
    --data-dir=/var/lib/etcd
9
 systemctl start etcd

1
# 查看可用版本
2
apt-cache madison kubeadm
3
kubeadm upgrade plan
4

5
# 升级控制平面
6
apt-get install -y kubeadm=1.28.0-*
7
kubeadm upgrade apply v1.28.0
8

9
# 升级节点
10
apt-get install -y kubelet=1.28.0-*
11
systemctl restart kubelet
12

13
# 升级插件
14
kubectl apply -f kube-proxy.yaml
15
kubectl rollout status daemonset/kube-proxy -n kube-system

1
# ResourceQuota：命名空间级别资源配额
2
apiVersion: v1
3
kind: ResourceQuota
4
metadata:
5
  name: compute-quota
6
  namespace: my-namespace
7
spec:
8
  hard:
9
    requests.cpu: "10"
10
    requests.memory: 20Gi
11
    limits.cpu: "20"
12
    limits.memory: 40Gi
13
    pods: "100"
14
    services: "20"
15

16
---
17

18
apiVersion: v1
19
kind: LimitRange
20
metadata:
21
  name: container-limits
22
  namespace: my-namespace
23
spec:
24
  limits:
25
    - max:
26
        cpu: "4"
27
        memory: 8Gi
28
      min:
29
        cpu: "100m"
30
        memory: 128Mi
31
      default:
32
        cpu: "500m"
33
        memory: 512Mi
34
      defaultRequest:
35
        cpu: "200m"
36
        memory: 256Mi
37
      type: Container

1
class ResourceCalculator:
2
    @staticmethod
3
    def calculate_node_capacity(node):
4
        """计算节点可用资源"""
5
        return {
6
            "cpu": node.status.capacity["cpu"],
7
            "memory": node.status.capacity["memory"],
8
        }
9

10
    @staticmethod
11
    def calculate_allocatable(node):
12
        """计算节点可分配资源"""
13
        capacity = ResourceCalculator.calculate_node_capacity(node)
14
        system_reserved = {
15
            "cpu": "500m",  # 系统预留
16
            "memory": "1Gi",
17
        }
18
        return {
19
            "cpu": capacity["cpu"] - parse_resources(system_reserved["cpu"]),
20
            "memory": capacity["memory"] - parse_resources(system_reserved["memory"]),
21
        }

1
# Fluent Bit 配置
2
apiVersion: v1
3
kind: ConfigMap
4
metadata:
5
  name: fluent-bit-config
6
  namespace: logging
7
data:
8
  fluent-bit.conf: |
9
    [SERVICE]
10
        Flush         5
11
        Log_Level     info
12

13
    [INPUT]
14
        Name              tail
15
        Path              /var/log/containers/*.log
16
        Parser            docker
17
        Tag               kube.*
18

19
    [FILTER]
20
        Name                kubernetes
21
        Match                kube.*
22
        Kube_URL            https://kubernetes.default.svc:443
23
        Kube_CA_File        /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
24
        Kube_Token_File     /var/run/secrets/kubernetes.io/serviceaccount/token
25

26
    [OUTPUT]
27
        Name              es
28
        Match               kube.*
29
        Host               elasticsearch.logging.svc
30
        Port               9200
31
        HTTP_User          elastic
32
        HTTP_Passwd        changeme
33
        Logstash_Format    On
34
        Logstash_Prefix    kubernetes

1
# 查看 Pod 日志
2
kubectl logs <pod-name> --previous  # 上一个容器
3
kubectl logs <pod-name> -c <container>  # 指定容器
4

5
# 实时跟踪日志
6
kubectl logs -f <pod-name>
7

8
# 查看带时间戳的日志
9
kubectl logs <pod-name> --timestamps
10

11
# 资源限制配置
12
apiVersion: v1
13
kind: Pod
14
metadata:
15
  name: logger
16
spec:
17
  containers:
18
  - name: logger
19
    image: busybox
20
    args:
21
    - /bin/sh
22
    - -c
23
    - |
24
        while true; do
25
          echo "$(date) - Log message" >> /var/log/app.log
26
          sleep 10
27
        done
28
    volumeMounts:
29
    - name: log
30
      mountPath: /var/log
31
  volumes:
32
  - name: log
33
    emptyDir: {}

1
# 快速诊断脚本示例
2
#!/bin/bash
3
echo "=== Cluster Status ==="
4
kubectl get nodes
5
kubectl get pods -A | grep -v Running | head -20
6

7
echo "=== Resource Usage ==="
8
kubectl top nodes
9
kubectl top pods -A | sort -k3 -rn | head -10
10

11
echo "=== Recent Events ==="
12
kubectl get events -A --sort-by='.lastTimestamp' | tail -20