ansible-devops/scripts/k8s-cluster-deploy.sh

#!/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'

# 命令记录存储（用于最后汇总展示）
declare -a EXECUTED_COMMANDS=()
LOG_FILE="/var/log/k8s-deploy-$(date +%Y%m%d-%H%M%S).log"

# 打印函数
info() { 
    echo -e "${YELLOW}[INFO]${NC} $1" | tee -a "$LOG_FILE"
}

success() { 
    echo -e "${GREEN}[SUCCESS]${NC} $1" | tee -a "$LOG_FILE"
}

error() { 
    echo -e "${RED}[ERROR]${NC} $1" | tee -a "$LOG_FILE"
    exit 1
}

step() { 
    echo -e "\n${BLUE}===== $1 =====${NC}" | tee -a "$LOG_FILE"
}

# 远程执行命令并记录
remote_exec() {
    local ip=$1
    local cmd=$2
    local desc=${3:-"执行命令"}
    
    info "在 $ip 上 $desc: ${YELLOW}$cmd${NC}"
    EXECUTED_COMMANDS+=("在 $ip 上: $cmd")
    
    # 执行远程命令并记录输出
    if ! ssh -o StrictHostKeyChecking=no -o ConnectTimeout=10 "$REMOTE_USER@$ip" "$cmd" >> "$LOG_FILE" 2>&1; then
        error "在 $ip 上执行命令失败: $cmd，请查看日志: $LOG_FILE"
    fi
}

# 固定路径（可根据实际调整）
KUBEADM_CONF="/opt/k8s-install-conf/kubeadm-conf.yaml"  # Master节点的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
CALICO_YAML_OFFICIAL="https://docs.projectcalico.org/v3.25/manifests/calico.yaml"
LOCAL_CALICO_YAML="/tmp/calico-v3.25-modified.yaml"
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 ==========================
MASTER_IPS=()  # 3个Master节点IP（顺序：Master1, Master2, Master3）
NODE_IPS=()    # Node节点IP列表

parse_args() {
    while [[ $# -gt 0 ]]; do
        case "$1" in
            --master-ips)
                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)
                IFS=',' read -ra IPS <<< "$2"
                NODE_IPS=("${IPS[@]}")
                shift 2
                ;;
            --help)
                echo "用法：$0 --master-ips <IP1,IP2,IP3> [--node-ips <IP1,IP2,...>]"
                echo "功能：部署K8s集群，包含3个Master节点和可选的Node节点"
                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Master IP
    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"
}


# ========================== 免密登录检测 ==========================
check_ssh_access() {
    step "前置检查：SSH免密登录验证"
    local all_ips=("${MASTER_IPS[@]}" "${NODE_IPS[@]}")
    local failed_ips=()

    for ip in "${all_ips[@]}"; do
        info "验证SSH免密登录：$REMOTE_USER@$ip"
        if ssh -o StrictHostKeyChecking=no -o ConnectTimeout=10 -o BatchMode=yes "$REMOTE_USER@$ip" "echo 'SSH连通正常'" >/dev/null 2>&1; then
            info "$ip SSH免密登录验证通过"
        else
            error "$ip SSH免密登录验证失败，请先执行: ssh-copy-id $REMOTE_USER@$ip 配置免密"
            failed_ips+=("$ip")
        fi
    done

    if [[ ${#failed_ips[@]} -gt 0 ]]; then
        error "以下节点SSH免密登录配置失败: ${failed_ips[*]}"
    fi
    
    success "所有节点SSH免密登录验证通过"
}


# ========================== 步骤1：Master1初始化 ==========================
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
    local init_cmd="kubeadm init --upload-certs --config $KUBEADM_CONF 2>&1"
    info "执行初始化命令：$init_cmd"
    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" "$init_cmd" | tee "$INIT_RESULT_FILE"
    local init_exit_code=${PIPESTATUS[0]}
    if [[ $init_exit_code -ne 0 ]]; then
        error "Master1初始化失败！查看详情：cat $INIT_RESULT_FILE"
    fi
    EXECUTED_COMMANDS+=("在 $MASTER1_IP 上初始化: $init_cmd")

    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 "${kube_cmds[@]}"; do
        remote_exec "$MASTER1_IP" "$cmd" "配置kubeconfig"
    done

    # 验证kubectl是否可用
    remote_exec "$MASTER1_IP" "kubectl get nodes" "验证kubectl"
    success "Master1 kubeconfig配置完成"
}


# ========================== 步骤3：提取join命令 ==========================
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"

    # 提取Master节点join命令
    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 "$master_join_cmd" ]] && error "未从初始化结果中提取到Master join命令"
    # 补充--config参数
    master_join_cmd="$master_join_cmd --config $KUBEADM_CONF"
    export MASTER_JOIN_CMD="$master_join_cmd"
    EXECUTED_COMMANDS+=("Master节点加入命令: $MASTER_JOIN_CMD")
    success "Master join命令提取完成："
    echo -e "${YELLOW}${MASTER_JOIN_CMD}${NC}"

    # 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"
    EXECUTED_COMMANDS+=("Node节点加入命令: $NODE_JOIN_CMD")
    success "Node join命令提取完成："
    echo -e "${YELLOW}${NODE_JOIN_CMD}${NC}"
}


# ========================== 步骤4：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
    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 $master_ip的配置文件不存在：$KUBEADM_CONF"
        fi

        # 远程执行join命令
        remote_exec "$master_ip" "$MASTER_JOIN_CMD" "执行加入集群命令"

        # 配置该Master的kubeconfig
        local kube_config_cmds=(
            "mkdir -p \$HOME/.kube"
            "cp -i /etc/kubernetes/admin.conf \$HOME/.kube/config"
            "chown \$(id -u):\$(id -g) \$HOME/.kube/config"
        )
        for cmd in "${kube_config_cmds[@]}"; do
            remote_exec "$master_ip" "$cmd" "配置kubeconfig"
        done

        success "Master $master_ip加入集群并配置完成"
    done

    # 验证所有Master节点状态
    remote_exec "$MASTER1_IP" "kubectl get nodes | grep master" "验证Master节点状态"
    success "所有Master节点加入完成"
}


# ========================== 步骤5：Node节点加入 ==========================
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
        local check_cmd="command -v kubeadm &>/dev/null && command -v kubelet &>/dev/null && systemctl is-active --quiet containerd"
        if ! ssh "$REMOTE_USER@$node_ip" "$check_cmd"; 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
        remote_exec "$node_ip" "$NODE_JOIN_CMD" "执行加入集群命令"

        success "Node $node_ip加入集群完成"
    done

    # 验证所有节点状态
    remote_exec "$MASTER1_IP" "kubectl get nodes" "验证所有节点状态"
    success "所有Node节点加入完成"
}


# ========================== 步骤6：部署Calico网络插件 ==========================
deploy_calico() {
    step "Step 6：部署Calico v3.25（华为云SWR镜像）"
    
    # 下载Calico 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）"

    # 替换YAML中的所有镜像地址为华为云SWR
    info "替换镜像地址为华为云SWR：$SWR_CALICO_PREFIX"
    sed -i "s#docker.io/calico/cni:v3.25.0#${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"

    # 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"
    EXECUTED_COMMANDS+=("上传Calico配置到 $MASTER1_IP: scp $LOCAL_CALICO_YAML $REMOTE_USER@$MASTER1_IP:$remote_calico_yaml")
    success "YAML上传完成"

    # 在Master1执行Calico部署
    local deploy_cmd="kubectl apply -f $remote_calico_yaml"
    remote_exec "$MASTER1_IP" "$deploy_cmd" "执行Calico部署"

    # 等待Calico Pod启动
    info "等待Calico Pod启动（最多5分钟）..."
    local wait_time=0
    while true; do
        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 [[ $wait_time -ge 300 ]]; then
            error "Calico Pod启动超时（5分钟），请手动检查：kubectl get pods -n calico-system"
        fi
        sleep 10
        wait_time=$((wait_time + 10))
        info "已等待${wait_time}秒，Calico Pod状态：$pod_status"
    done

    # 验证Calico状态
    remote_exec "$MASTER1_IP" "kubectl get pods -n calico-system" "验证Calico状态"
    remote_exec "$MASTER1_IP" "kubectl get nodes -o wide | grep -E 'STATUS|Ready'" "验证节点网络状态"
    success "Calico v3.25（华为云SWR）部署完成，集群网络已就绪"

    # 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"
}


# ========================== 步骤7：展示执行的所有命令 ==========================
show_executed_commands() {
    step "Step 7：执行命令汇总"
    echo -e "${YELLOW}以下是部署过程中执行的关键命令：${NC}"
    echo -e "----------------------------------------------------------------------"
    local count=1
    for cmd in "${EXECUTED_COMMANDS[@]}"; do
        echo -e "${count}. ${cmd}"
        ((count++))
    done
    echo -e "----------------------------------------------------------------------"
}


# ========================== 主流程 ==========================
main() {
    # 记录开始时间
    local start_time=$(date +%s)
    
    # 1. 解析参数
    parse_args "$@"
    
    # 2. 检查SSH免密登录
    check_ssh_access
    
    # 3. 执行核心步骤
    init_master1          
    config_master1_kubeconfig  
    extract_join_commands 
    join_other_masters    
    join_nodes            
    deploy_calico         
    
    # 4. 展示执行的命令
    show_executed_commands
    
    # 5. 集群部署完成总结
    step "集群部署完成总结"
    local end_time=$(date +%s)
    local duration=$((end_time - start_time))
    echo -e "${GREEN}========================================"
    echo -e "K8s集群部署全流程完成！"
    echo -e "部署耗时：$((duration/60))分$((duration%60))秒"
    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：$INIT_RESULT_FILE"
    echo -e "  5. 部署日志：$LOG_FILE"
    echo -e "========================================"
}

# 启动主流程
main "$@"