欢迎您访问程序员文章站本站旨在为大家提供分享程序员计算机编程知识!
您现在的位置是: 首页

etcd集群之k8s statefulset实现

程序员文章站 2024-03-11 16:28:13
...

说明

etcd用k8s的statefulset实现, 一个statefulset的yaml就能实现etcd的集群的动态伸缩, 当你减少statefulset的replicas的时候(或者节点出故障down掉), 自动的把节点从etcd集群中移除,当时增加statefulset的replicas的时候(或者故障节点恢复),自动的把节点添加到etcd集群中. 但是当你减少的集群节点数目小于初始化集群的时候设置的最小集群节点个数的时候, 减少集群节点(或者节点出故障down掉)不会将节点从集群中移除, 新增节点(或者故障节点恢复),自动的把新增节点(或者故障节点恢复)更新到etcd集群中. 声明下该statefulset是基于k8s官方提供的原版yaml修改测试后的. 文件位置,最新源码中(老版本源码中不一定会有): https://github.com/kubernetes/kubernetes/test/e2e/testing-manifests/statefulset/etcd 中的yaml. 但是该yaml无法正常的创建etcd集群, 存在很多问题. 可以自己去测试.

修改后的yaml如下. 如果自己的k8s中不提供存储支持, 请自行将yaml的的pvc改成emptyDir的volume即可. 另外镜像可以自己换成etcd官方镜像.

yaml如下:

apiVersion: v1
kind: Service
metadata:
  name: etcd
  labels:
    app: etcd
spec:
  ports:
  - port: 2380
    name: etcd-server
  - port: 2379
    name: etcd-client
  clusterIP: None
  selector:
    app: etcd
  publishNotReadyAddresses: true

--- 
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: etcd
  labels:
    app: etcd
spec:
  serviceName: etcd
  replicas: 3
  selector:
    matchLabels:
      app: etcd
  template:
    metadata:
      name: etcd
      labels:
        app: etcd
    spec:
      containers:
      - name: etcd
        image: gcr.k8s.io/etcd:3.2.24
        imagePullPolicy: Always
        ports:
        - containerPort: 2380
          name: peer
        - containerPort: 2379
          name: client
        resources:
          requests:
            cpu: 100m
            memory: 512Mi
        env:
        - name: INITIAL_CLUSTER_SIZE
          value: "3"
        - name: SET_NAME
          value: etcd
        - name: NAMESPACE
          valueFrom: 
            fieldRef: 
              fieldPath: meta.namepace
        volumeMounts:
        - name: datadir
          mountPath: /var/run/etcd
        lifecycle:
          preStop:
            exec:
              command:
                - "/bin/sh"
                - "-ec"
                - |
                  EPS=""
                  for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                      EPS="${EPS}${EPS:+,}http://${SET_NAME}-${i}.${SET_NAME}:2379"
                  done

                  HOSTNAME=$(hostname)

                  member_hash() {
                      etcdctl member list | grep http://${HOSTNAME}.${SET_NAME}:2380 | cut -d':' -f1 | cut -d'[' -f1
                  }

                  # Remove everything otherwise the cluster will no longer scale-up
                  SET_ID=${HOSTNAME:5:${#HOSTNAME}}
                  # adding a new member to existing cluster (assuming all initial pods are available)
                  if [ "${SET_ID}" -ge ${INITIAL_CLUSTER_SIZE} ]; then
                    echo "Removing ${HOSTNAME} from etcd cluster"
                    ETCDCTL_ENDPOINT=${EPS} etcdctl member remove $(member_hash)
                    if [ $? -eq 0 ]; then
                      rm -rf /var/run/etcd/*
                    fi
                  fi
                  
        command:
          - "/bin/sh"
          - "-ec"
          - |
            HOSTNAME=$(hostname)

            # store member id into PVC for later member replacement
            collect_member() {
                while ! etcdctl member list &>/dev/null; do sleep 1; done
                etcdctl member list | grep http://${HOSTNAME}.${SET_NAME}:2380 | cut -d':' -f1 | cut -d'[' -f1 > /var/run/etcd/member_id
                exit 0
            }

            eps() {
                EPS=""
                for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                    EPS="${EPS}${EPS:+,}http://${SET_NAME}-${i}.${SET_NAME}:2379"
                done
                echo ${EPS}
            }

            member_hash() {
                etcdctl member list | grep http://${HOSTNAME}.${SET_NAME}:2380 | cut -d':' -f1 | cut -d'[' -f1
            }

            # re-joining after failure?
            if [ -e /var/run/etcd/default.etcd ]; then
                echo "Re-joining etcd member"
                member_id=$(cat /var/run/etcd/member_id)

                # re-join member
                POD_IP=$(hostname -i)
                ETCDCTL_ENDPOINT=$(eps) etcdctl member update ${member_id} http://${HOSTNAME}.${SET_NAME}:2380
                exec etcd --name ${HOSTNAME} \
                    --listen-peer-urls http://${POD_IP}:2380 \
                    --listen-client-urls http://${POD_IP}:2379,http://127.0.0.1:2379 \
                    --advertise-client-urls http://${POD_IP}:2379 \
                    --data-dir /var/run/etcd/default.etcd
            fi

            # etcd-SET_ID

            SET_ID=${HOSTNAME:5:${#HOSTNAME}}
            
            # adding a new member to existing cluster (assuming all initial pods are available)
            if [ "${SET_ID}" -ge ${INITIAL_CLUSTER_SIZE} ]; then
                export ETCDCTL_ENDPOINT=$(eps)

                # member already added?
                MEMBER_HASH=$(member_hash)


                if [ -n "${MEMBER_HASH}" ]; then
                    # the member hash exists but for some reason etcd failed
                    # as the datadir has not be created, we can remove the member
                    # and retrieve new hash
                    etcdctl member remove ${MEMBER_HASH}
                fi

                echo "Adding new member"
                etcdctl member add ${HOSTNAME} http://${HOSTNAME}.${SET_NAME}:2380 | grep "^ETCD_" > /var/run/etcd/new_member_envs

                if [ $? -ne 0 ]; then
                    echo "Exiting"
                    rm -f /var/run/etcd/new_member_envs
                    exit 1
                fi

                cat /var/run/etcd/new_member_envs
                source /var/run/etcd/new_member_envs

                collect_member &

                POD_IP=$(hostname -i)
                exec etcd --name ${HOSTNAME} \
                    --listen-peer-urls http://${POD_IP}:2380 \
                    --listen-client-urls http://${POD_IP}:2379,http://127.0.0.1:2379 \
                    --advertise-client-urls http://${POD_IP}:2379 \
                    --data-dir /var/run/etcd/default.etcd \
                    --initial-advertise-peer-urls http://${HOSTNAME}.${SET_NAME}:2380 \
                    --initial-cluster ${ETCD_INITIAL_CLUSTER} \
                    --initial-cluster-state ${ETCD_INITIAL_CLUSTER_STATE}
            fi

            for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                while true; do
                    echo "Waiting for ${SET_NAME}-${i}.${SET_NAME} to come up"
                    ping -W 1 -c 1 ${SET_NAME}-${i}.${SET_NAME}.${NAMESPACE}.svc.cluster.local > /dev/null && break
                    sleep 1s
                done
            done

            PEERS=""
            for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                PEERS="${PEERS}${PEERS:+,}${SET_NAME}-${i}=http://${SET_NAME}-${i}.${SET_NAME}:2380"
            done
            
            collect_member &

            # join member
            POD_IP=$(hostname -i)
            exec etcd --name ${HOSTNAME} \
                --initial-advertise-peer-urls http://${POD_IP}:2380 \
                --listen-peer-urls http://${POD_IP}:2380 \
                --listen-client-urls http://${POD_IP}:2379,http://127.0.0.1:2379 \
                --advertise-client-urls http://${POD_IP}:2379 \
                --initial-cluster-token etcd-cluster-1 \
                --initial-cluster ${PEERS} \
                --initial-cluster-state new \
                --data-dir /var/run/etcd/default.etcd
  volumeClaimTemplates:
  - metadata:
      name: datadir
    spec:
      accessModes:
        - "ReadWriteOnce"
      resources:
        requests:
          # upstream recommended max is 700M
          storage: 1Gi

emptyDir的yaml如下:

apiVersion: v1
kind: Service
metadata:
  name: etcd
  labels:
    app: etcd
spec:
  ports:
  - port: 2380
    name: etcd-server
  - port: 2379
    name: etcd-client
  clusterIP: None
  selector:
    app: etcd
  publishNotReadyAddresses: true
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: etcd
  labels:
    app: etcd
spec:
  serviceName: etcd
  replicas: 3
  selector:
    matchLabels:
      app: etcd
  template:
    metadata:
      name: etcd
      labels:
        app: etcd
    spec:
      volumes:
      - name: datadir
        emptyDir: {}
      containers:
      - name: etcd
        image: gcr.k8s.io/etcd:3.2.24
        imagePullPolicy: Always
        ports:
        - containerPort: 2380
          name: peer
        - containerPort: 2379
          name: client
        resources:
          requests:
            cpu: 100m
            memory: 512Mi
        env:
        - name: INITIAL_CLUSTER_SIZE
          value: "3"
        - name: SET_NAME
          value: etcd
        - name: NAMESPACE
          valueFrom: 
            fieldRef: 
              fieldPath: meta.namepace
        volumeMounts:
        - name: datadir
          mountPath: /var/run/etcd
        lifecycle:
          preStop:
            exec:
              command:
                - "/bin/sh"
                - "-ec"
                - |
                  EPS=""
                  for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                      EPS="${EPS}${EPS:+,}http://${SET_NAME}-${i}.${SET_NAME}:2379"
                  done

                  HOSTNAME=$(hostname)

                  member_hash() {
                      etcdctl member list | grep http://${HOSTNAME}.${SET_NAME}:2380 | cut -d':' -f1 | cut -d'[' -f1
                  }

                  # Remove everything otherwise the cluster will no longer scale-up
                  SET_ID=${HOSTNAME:5:${#HOSTNAME}}
                  # adding a new member to existing cluster (assuming all initial pods are available)
                  if [ "${SET_ID}" -ge ${INITIAL_CLUSTER_SIZE} ]; then
                    echo "Removing ${HOSTNAME} from etcd cluster"
                    ETCDCTL_ENDPOINT=${EPS} etcdctl member remove $(member_hash)
                    if [ $? -eq 0 ]; then
                      rm -rf /var/run/etcd/*
                    fi
                  fi
                  
        command:
          - "/bin/sh"
          - "-ec"
          - |
            HOSTNAME=$(hostname)

            # store member id into PVC for later member replacement
            collect_member() {
                while ! etcdctl member list &>/dev/null; do sleep 1; done
                etcdctl member list | grep http://${HOSTNAME}.${SET_NAME}:2380 | cut -d':' -f1 | cut -d'[' -f1 > /var/run/etcd/member_id
                exit 0
            }

            eps() {
                EPS=""
                for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                    EPS="${EPS}${EPS:+,}http://${SET_NAME}-${i}.${SET_NAME}:2379"
                done
                echo ${EPS}
            }

            member_hash() {
                etcdctl member list | grep http://${HOSTNAME}.${SET_NAME}:2380 | cut -d':' -f1 | cut -d'[' -f1
            }

            # re-joining after failure?
            if [ -e /var/run/etcd/default.etcd ]; then
                echo "Re-joining etcd member"
                member_id=$(cat /var/run/etcd/member_id)

                # re-join member
                POD_IP=$(hostname -i)
                ETCDCTL_ENDPOINT=$(eps) etcdctl member update ${member_id} http://${HOSTNAME}.${SET_NAME}:2380
                exec etcd --name ${HOSTNAME} \
                    --listen-peer-urls http://${POD_IP}:2380 \
                    --listen-client-urls http://${POD_IP}:2379,http://127.0.0.1:2379 \
                    --advertise-client-urls http://${POD_IP}:2379 \
                    --data-dir /var/run/etcd/default.etcd
            fi

            # etcd-SET_ID

            SET_ID=${HOSTNAME:5:${#HOSTNAME}}
            
            # adding a new member to existing cluster (assuming all initial pods are available)
            if [ "${SET_ID}" -ge ${INITIAL_CLUSTER_SIZE} ]; then
                export ETCDCTL_ENDPOINT=$(eps)

                # member already added?
                MEMBER_HASH=$(member_hash)


                if [ -n "${MEMBER_HASH}" ]; then
                    # the member hash exists but for some reason etcd failed
                    # as the datadir has not be created, we can remove the member
                    # and retrieve new hash
                    etcdctl member remove ${MEMBER_HASH}
                fi

                echo "Adding new member"
                etcdctl member add ${HOSTNAME} http://${HOSTNAME}.${SET_NAME}:2380 | grep "^ETCD_" > /var/run/etcd/new_member_envs

                if [ $? -ne 0 ]; then
                    echo "Exiting"
                    rm -f /var/run/etcd/new_member_envs
                    exit 1
                fi

                cat /var/run/etcd/new_member_envs
                source /var/run/etcd/new_member_envs

                collect_member &

                POD_IP=$(hostname -i)
                exec etcd --name ${HOSTNAME} \
                    --listen-peer-urls http://${POD_IP}:2380 \
                    --listen-client-urls http://${POD_IP}:2379,http://127.0.0.1:2379 \
                    --advertise-client-urls http://${POD_IP}:2379 \
                    --data-dir /var/run/etcd/default.etcd \
                    --initial-advertise-peer-urls http://${HOSTNAME}.${SET_NAME}:2380 \
                    --initial-cluster ${ETCD_INITIAL_CLUSTER} \
                    --initial-cluster-state ${ETCD_INITIAL_CLUSTER_STATE}
            fi

            for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                while true; do
                    echo "Waiting for ${SET_NAME}-${i}.${SET_NAME} to come up"
                    ping -W 1 -c 1 ${SET_NAME}-${i}.${SET_NAME}.${NAMESPACE}.svc.cluster.local > /dev/null && break
                    sleep 1s
                done
            done

            PEERS=""
            for i in $(seq 0 $((${INITIAL_CLUSTER_SIZE} - 1))); do
                PEERS="${PEERS}${PEERS:+,}${SET_NAME}-${i}=http://${SET_NAME}-${i}.${SET_NAME}:2380"
            done
            
            collect_member &

            # join member
            POD_IP=$(hostname -i)
            exec etcd --name ${HOSTNAME} \
                --initial-advertise-peer-urls http://${POD_IP}:2380 \
                --listen-peer-urls http://${POD_IP}:2380 \
                --listen-client-urls http://${POD_IP}:2379,http://127.0.0.1:2379 \
                --advertise-client-urls http://${POD_IP}:2379 \
                --initial-cluster-token etcd-cluster-1 \
                --initial-cluster ${PEERS} \
                --initial-cluster-state new \
                --data-dir /var/run/etcd/default.etcd
  # volumeClaimTemplates:
  # - metadata:
  #     name: datadir
  #   spec:
  #     accessModes:
  #       - "ReadWriteOnce"
  #     resources:
  #       requests:
  #         # upstream recommended max is 700M
  #         storage: 1Gi