From e15fd8b6f5f4db643a2fe98285b8b8cb095fa56f Mon Sep 17 00:00:00 2001 From: joy Date: Thu, 30 Oct 2025 18:26:21 +0800 Subject: [PATCH] =?UTF-8?q?=E6=B7=BB=E5=8A=A0=20scripts/k8s-cluster-deploy?= =?UTF-8?q?.sh?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit aaa --- scripts/k8s-cluster-deploy.sh | 340 ++++++++++++++++++++++++++++++++++ 1 file changed, 340 insertions(+) create mode 100644 scripts/k8s-cluster-deploy.sh diff --git a/scripts/k8s-cluster-deploy.sh b/scripts/k8s-cluster-deploy.sh new file mode 100644 index 0000000..39be23e --- /dev/null +++ b/scripts/k8s-cluster-deploy.sh @@ -0,0 +1,340 @@ +#!/bin/bash +set -euo pipefail + +# ========================== 全局配置与工具函数 ========================== +# 颜色定义 +RED='\033[0;31m' +GREEN='\033[0;32m' +YELLOW='\033[1;33m' +BLUE='\033[0;34m' +NC='\033[0m' + +# 打印函数 +info() { echo -e "${YELLOW}[INFO]${NC} $1"; } +success() { echo -e "${GREEN}[SUCCESS]${NC} $1"; } +error() { echo -e "${RED}[ERROR]${NC} $1"; exit 1; } +step() { echo -e "\n${BLUE}===== $1 =====${NC}"; } + +# 固定路径（可根据实际调整） +KUBEADM_CONF="/opt/k8s-install-conf/kubeadm-conf.yaml" # Master1的kubeadm配置文件 +INIT_RESULT_FILE="/opt/k8s-install-conf/kubeadm-init-result.txt" # 初始化结果保存文件 +# 关键修改1：Calico v3.25官方YAML地址（替换原v3.26） +CALICO_YAML_OFFICIAL="https://docs.projectcalico.org/v3.25/manifests/calico.yaml" +# 本地临时Calico YAML路径（用于镜像替换） +LOCAL_CALICO_YAML="/tmp/calico-v3.25-modified.yaml" +# 关键修改2：华为云SWR镜像前缀（替换原docker.io/calico） +SWR_CALICO_PREFIX="swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/calico" +REMOTE_USER="root" # 远程节点登录用户 + + +# ========================== 参数解析（核心：指定Master/Node IP） ========================== +# 初始化参数 +MASTER_IPS=() # 3个Master节点IP（顺序：Master1, Master2, Master3） +NODE_IPS=() # Node节点IP列表（支持多个，用逗号分隔） + +parse_args() { + while [[ $# -gt 0 ]]; do + case "$1" in + --master-ips) + # 解析3个Master IP（格式："192.168.61.10,192.168.61.11,192.168.61.12"） + IFS=',' read -ra IPS <<< "$2" + if [[ ${#IPS[@]} -ne 3 ]]; then + error "--master-ips必须包含3个IP（逗号分隔），当前提供${#IPS[@]}个" + fi + MASTER_IPS=("${IPS[@]}") + shift 2 + ;; + --node-ips) + # 解析Node IP列表（格式："192.168.61.20,192.168.61.21"） + IFS=',' read -ra IPS <<< "$2" + NODE_IPS=("${IPS[@]}") + shift 2 + ;; + --help) + echo "用法：$0 --master-ips [--node-ips ]" + echo "功能：Master1初始化→远程Master/Node加入→Calico(v3.25+华为云SWR)部署" + echo "参数说明：" + echo " --master-ips 必选，3个Master节点IP（顺序：Master1(初始化节点), Master2, Master3）" + echo " --node-ips 可选，Node节点IP列表（逗号分隔，如不指定则仅部署Master集群）" + echo "示例：" + echo " $0 --master-ips 192.168.61.10,192.168.61.11,192.168.61.12 --node-ips 192.168.61.20,192.168.61.21" + exit 0 + ;; + *) + error "未知参数：$1（执行$0 --help查看用法）" + ;; + esac + done + + # 校验必选参数 + if [[ ${#MASTER_IPS[@]} -eq 0 ]]; then + error "缺少必选参数--master-ips（需指定3个Master节点IP）" + fi + + # 提取Master1 IP（初始化节点） + export MASTER1_IP="${MASTER_IPS[0]}" + export MASTER2_IP="${MASTER_IPS[1]}" + export MASTER3_IP="${MASTER_IPS[2]}" + info "已识别节点：" + info "Master1（初始化）：$MASTER1_IP" + info "Master2：$MASTER2_IP" + info "Master3：$MASTER3_IP" + [[ ${#NODE_IPS[@]} -gt 0 ]] && info "Node节点：${NODE_IPS[*]}" || info "未指定Node节点，仅部署Master集群" +} + + +# ========================== 步骤1：Master1初始化（核心） ========================== +init_master1() { + step "Step 1：Master1节点初始化（$MASTER1_IP）" + + # 检查Master1的kubeadm配置文件是否存在 + if ! ssh "$REMOTE_USER@$MASTER1_IP" "test -f $KUBEADM_CONF"; then + error "Master1的kubeadm配置文件不存在：$MASTER1_IP:$KUBEADM_CONF（请先执行配置分发脚本）" + fi + + # 执行kubeadm init（带证书上传，支持其他Master加入） + info "执行初始化命令：kubeadm init --upload-certs --config $KUBEADM_CONF" + info "初始化结果将保存到：$INIT_RESULT_FILE" + + # 远程执行初始化，同时保存输出到本地文件 + ssh "$REMOTE_USER@$MASTER1_IP" "kubeadm init --upload-certs --config $KUBEADM_CONF 2>&1" | tee "$INIT_RESULT_FILE" + local init_exit_code=${PIPESTATUS[0]} # 获取远程命令退出码 + if [[ $init_exit_code -ne 0 ]]; then + error "Master1初始化失败！查看详情：cat $INIT_RESULT_FILE" + fi + + success "Master1初始化完成，结果已保存到：$INIT_RESULT_FILE" + info "初始化输出预览（前20行）：" + head -n 20 "$INIT_RESULT_FILE" +} + + +# ========================== 步骤2：Master1配置kubeconfig ========================== +config_master1_kubeconfig() { + step "Step 2：配置Master1的kubeconfig" + + # 远程执行kubeconfig配置命令 + local kube_cmds=( + "mkdir -p \$HOME/.kube" + "cp -i /etc/kubernetes/admin.conf \$HOME/.kube/config" + "chown \$(id -u):\$(id -g) \$HOME/.kube/config" + "echo 'export KUBECONFIG=/etc/kubernetes/admin.conf' >> ~/.bash_profile" + "source ~/.bash_profile" + ) + + for cmd in "${kube_cmds[@]}"; do + info "在Master1执行：$cmd" + ssh "$REMOTE_USER@$MASTER1_IP" "$cmd" || error "执行命令失败：$cmd" + done + + # 验证kubectl是否可用 + info "验证kubectl：在Master1执行 kubectl get nodes" + ssh "$REMOTE_USER@$MASTER1_IP" "kubectl get nodes" || error "kubectl配置失败，无法获取节点信息" + success "Master1 kubeconfig配置完成" +} + + +# ========================== 步骤3：提取join命令（从初始化结果中） ========================== +extract_join_commands() { + step "Step 3：从初始化结果中提取join命令" + + # 检查结果文件是否存在 + [[ -f "$INIT_RESULT_FILE" ]] || error "初始化结果文件不存在：$INIT_RESULT_FILE" + + # 1. 提取Master节点join命令（含--control-plane和--certificate-key） + info "提取Master节点join命令" + local master_join_cmd + master_join_cmd=$(grep -A 3 "You can now join any number of control-plane nodes by running the following command on each as root" "$INIT_RESULT_FILE" | grep -v "You can now" | tr -d '\n' | sed 's/ */ /g') + [[ -z "$master_join_cmd" ]] && error "未从初始化结果中提取到Master join命令" + # 补充--config参数（其他Master也需用自己的配置文件） + master_join_cmd="$master_join_cmd --config $KUBEADM_CONF" + export MASTER_JOIN_CMD="$master_join_cmd" + success "Master join命令提取完成：" + echo -e "${YELLOW}${MASTER_JOIN_CMD}${NC}" + + # 2. 提取Node节点join命令 + info "提取Node节点join命令" + local node_join_cmd + node_join_cmd=$(grep -A 2 "Then you can join any number of worker nodes by running the following on each as root" "$INIT_RESULT_FILE" | grep -v "Then you can" | tr -d '\n' | sed 's/ */ /g') + [[ -z "$node_join_cmd" ]] && error "未从初始化结果中提取到Node join命令" + export NODE_JOIN_CMD="$node_join_cmd" + success "Node join命令提取完成：" + echo -e "${YELLOW}${NODE_JOIN_CMD}${NC}" +} + + +# ========================== 步骤4：远程其他Master节点加入 ========================== +join_other_masters() { + step "Step 4：远程Master2（$MASTER2_IP）和Master3（$MASTER3_IP）加入集群" + + # 定义要加入的Master节点列表（排除Master1） + local other_masters=("$MASTER2_IP" "$MASTER3_IP") + + for master_ip in "${other_masters[@]}"; do + info "处理Master节点：$master_ip" + + # 检查该Master的kubeadm配置文件是否存在 + if ! ssh "$REMOTE_USER@$master_ip" "test -f $KUBEADM_CONF"; then + error "Master $master_ip的kubeconfig配置文件不存在：$KUBEADM_CONF" + fi + + # 远程执行join命令 + info "在$master_ip执行join命令：$MASTER_JOIN_CMD" + ssh "$REMOTE_USER@$master_ip" "$MASTER_JOIN_CMD" || error "Master $master_ip加入失败" + + # 配置该Master的kubeconfig（否则无法本地执行kubectl） + info "配置$master_ip的kubeconfig" + ssh "$REMOTE_USER@$master_ip" "mkdir -p \$HOME/.kube && cp -i /etc/kubernetes/admin.conf \$HOME/.kube/config && chown \$(id -u):\$(id -g) \$HOME/.kube/config" + + success "Master $master_ip加入集群并配置完成" + done + + # 验证所有Master节点状态 + info "验证所有Master节点状态（从Master1执行）" + ssh "$REMOTE_USER@$MASTER1_IP" "kubectl get nodes | grep master" || error "Master节点状态验证失败" + success "所有Master节点加入完成" +} + + +# ========================== 步骤5：远程Node节点加入（可选） ========================== +join_nodes() { + if [[ ${#NODE_IPS[@]} -eq 0 ]]; then + info "未指定Node节点，跳过Node加入步骤" + return + fi + + step "Step 5：远程Node节点加入集群（共${#NODE_IPS[@]}个）" + + for node_ip in "${NODE_IPS[@]}"; do + info "处理Node节点：$node_ip" + + # 检查Node节点是否已安装基础组件（kubeadm、kubelet、containerd） + if ! ssh "$REMOTE_USER@$node_ip" "command -v kubeadm &>/dev/null && command -v kubelet &>/dev/null && systemctl is-active --quiet containerd"; then + error "Node $node_ip未安装基础组件（需先安装kubeadm、kubelet、containerd）" + fi + + # 远程执行Node join命令 + info "在$node_ip执行join命令：$NODE_JOIN_CMD" + ssh "$REMOTE_USER@$node_ip" "$NODE_JOIN_CMD" || error "Node $node_ip加入失败" + + success "Node $node_ip加入集群完成" + done + + # 验证所有节点状态 + info "验证所有节点状态（从Master1执行）" + ssh "$REMOTE_USER@$MASTER1_IP" "kubectl get nodes" || error "节点状态验证失败" + success "所有Node节点加入完成" +} + + +# ========================== 步骤6：部署Calico网络插件（v3.25+华为云SWR） ========================== +deploy_calico() { + step "Step 6：部署Calico v3.25（华为云SWR镜像）" + + # 子步骤1：下载Calico v3.25官方YAML到本地 + info "下载Calico v3.25官方YAML：$CALICO_YAML_OFFICIAL" + if command -v wget &>/dev/null; then + wget -q -O "$LOCAL_CALICO_YAML" "$CALICO_YAML_OFFICIAL" || error "wget下载Calico YAML失败" + elif command -v curl &>/dev/null; then + curl -s -o "$LOCAL_CALICO_YAML" "$CALICO_YAML_OFFICIAL" || error "curl下载Calico YAML失败" + else + error "请先安装wget或curl" + fi + [[ -f "$LOCAL_CALICO_YAML" ]] || error "Calico YAML下载失败，文件不存在" + success "Calico YAML下载完成（本地路径：$LOCAL_CALICO_YAML）" + + # 子步骤2：替换YAML中的所有镜像地址为华为云SWR（核心修改） + info "替换镜像地址为华为云SWR：$SWR_CALICO_PREFIX" + # 替换3个核心镜像（cni、node、kube-controllers，均为v3.25.0版本） + sed -i "s#docker.io/calico/cni:v3.25.0#${SWR_CALICO_PREFIX}/cni:v3.25.0#g" "$LOCAL_CALICO_YAML" + sed -i "s#docker.io/calico/node:v3.25.0#${SWR_CALICO_PREFIX}/node:v3.25.0#g" "$LOCAL_CALICO_YAML" + sed -i "s#docker.io/calico/kube-controllers:v3.25.0#${SWR_CALICO_PREFIX}/kube-controllers:v3.25.0#g" "$LOCAL_CALICO_YAML" + + # 验证替换结果 + info "验证镜像替换结果（查看前10行含镜像的内容）" + grep -n "image:" "$LOCAL_CALICO_YAML" | head -n 10 || error "Calico YAML中未找到镜像配置，替换失败" + success "Calico镜像地址替换完成" + + # 子步骤3：将修改后的YAML上传到Master1节点 + local remote_calico_yaml="/opt/k8s-install-conf/calico-v3.25-swr.yaml" + info "上传修改后的YAML到Master1：$remote_calico_yaml" + scp -o StrictHostKeyChecking=no "$LOCAL_CALICO_YAML" "$REMOTE_USER@$MASTER1_IP:$remote_calico_yaml" || error "YAML上传失败" + success "YAML上传完成" + + # 子步骤4：在Master1执行Calico部署 + info "执行部署命令：kubectl apply -f $remote_calico_yaml" + ssh "$REMOTE_USER@$MASTER1_IP" "kubectl apply -f $remote_calico_yaml" || error "Calico部署失败" + + # 子步骤5：等待Calico Pod启动（最多等待5分钟） + info "等待Calico Pod启动（最多5分钟）..." + local wait_time=0 + while true; do + # 检查calico-system命名空间下所有Pod是否Running + local pod_status + pod_status=$(ssh "$REMOTE_USER@$MASTER1_IP" "kubectl get pods -n calico-system -o jsonpath='{.items[*].status.phase}' 2>/dev/null") + if [[ "$pod_status" =~ ^(Running\ )*Running$ && -n "$pod_status" ]]; then + break + fi + if [[ $wait_time -ge 300 ]]; then # 5分钟=300秒 + error "Calico Pod启动超时（5分钟），请手动检查：kubectl get pods -n calico-system" + fi + sleep 10 + wait_time=$((wait_time + 10)) + info "已等待${wait_time}秒，Calico Pod状态：$pod_status" + done + + # 子步骤6：验证Calico状态 + info "验证Calico状态（从Master1执行）" + ssh "$REMOTE_USER@$MASTER1_IP" "kubectl get pods -n calico-system" || error "Calico状态验证失败" + # 额外验证网络插件是否就绪 + ssh "$REMOTE_USER@$MASTER1_IP" "kubectl get nodes -o wide | grep -E 'STATUS|Ready'" || error "节点网络未就绪" + success "Calico v3.25（华为云SWR）部署完成，集群网络已就绪" + + # 子步骤7：清理本地临时YAML文件 + info "清理本地临时文件：$LOCAL_CALICO_YAML" + rm -f "$LOCAL_CALICO_YAML" || info "本地文件清理失败，可手动删除" +} + + +# ========================== 主流程：串联所有步骤 ========================== +main() { + # 1. 解析参数 + parse_args "$@" + + # 2. 前置检查：确保SSH免密登录正常 + step "前置检查：SSH免密登录验证" + local all_ips=("${MASTER_IPS[@]}" "${NODE_IPS[@]}") + for ip in "${all_ips[@]}"; do + info "验证SSH免密登录：$REMOTE_USER@$ip" + if ! ssh -o StrictHostKeyChecking=no "$REMOTE_USER@$ip" "echo 'SSH连通正常' >/dev/null"; then + error "SSH免密登录失败：$ip（请先执行 ssh-copy-id $REMOTE_USER@$ip 配置免密）" + fi + done + success "所有节点SSH免密登录验证通过" + + # 3. 执行核心步骤 + init_master1 # Master1初始化 + config_master1_kubeconfig # Master1配置kubeconfig + extract_join_commands # 提取join命令 + join_other_masters # 其他Master加入 + join_nodes # Node加入（可选） + deploy_calico # 部署Calico（v3.25+SWR） + + # 4. 集群部署完成总结 + step "集群部署完成总结" + echo -e "${GREEN}========================================" + echo -e "K8s集群部署全流程完成！" + echo -e "关键信息：" + echo -e " 1. Calico版本：v3.25（镜像源：华为云SWR）" + echo -e " 2. 集群节点状态：ssh $REMOTE_USER@$MASTER1_IP 'kubectl get nodes'" + echo -e " 3. Calico状态：ssh $REMOTE_USER@$MASTER1_IP 'kubectl get pods -n calico-system'" + echo -e " 4. 初始化结果：$INIT_RESULT_FILE" + echo -e "========================================" +} + +# 启动主流程 +main "$@" +chmod +x k8s-cluster-deploy.sh + +# 部署3个Master + 2个Node（替换为你的IP） +#bash k8s-cluster-deploy.sh --master-ips 192.168.61.10,192.168.61.11,192.168.61.12 --node-ips 192.168.61.20,192.168.61.21