一、为什么需要高可用架构

单点故障是Web服务的最大隐患。当服务器宕机、网络中断或应用崩溃时，没有高可用架构的网站将完全不可用。ThinkPHP作为国内主流PHP框架，其高可用部署需要从负载均衡、数据库集群、会话共享、文件同步等多个维度进行设计。

高可用核心目标：
– 零停机时间：故障自动转移，用户无感知
– 水平扩展：业务增长时快速增加节点
– 数据安全：数据库主从复制，数据零丢失
– 故障自愈：自动检测并恢复异常节点

二、整体架构设计

                    ┌─────────────┐
                    │   Nginx LB  │
                    │ (主负载均衡) │
                    └──────┬──────┘
                           │
              ┌────────────┼────────────┐
              │            │            │
        ┌─────┴─────┐ ┌───┴────┐ ┌────┴─────┐
        │  Node 1   │ │ Node 2 │ │  Node 3  │
        │  ThinkPHP │ │ThinkPHP│ │ ThinkPHP │
        │  + PHP-FPM│ │+FPM    │ │  + FPM   │
        └─────┬─────┘ └───┬────┘ └────┬─────┘
              │            │            │
        ┌─────┴────────────┴────────────┴─────┐
        │            共享存储层                  │
        │  Redis(会话) + NFS(文件) + MariaDB   │
        └─────────────────────────────────────┘

三、环境准备

3.1 服务器规划

节点	IP地址	用途	配置建议
LB-1	192.168.1.100	Nginx主负载均衡	2核2G
LB-2	192.168.1.101	Nginx备用负载均衡	2核2G
APP-1	192.168.1.110	ThinkPHP应用节点1	4核4G
APP-2	192.168.1.111	ThinkPHP应用节点2	4核4G
DB-M	192.168.1.120	MariaDB主库	4核8G SSD
DB-S	192.168.1.121	MariaDB从库	4核8G SSD
Redis	192.168.1.130	Redis会话/缓存	2核4G

3.2 基础软件安装

所有应用节点执行：

# 更新系统
sudo apt update && sudo apt upgrade -y

# 安装Nginx、PHP和扩展
sudo apt install -y nginx php8.3-fpm php8.3-cli php8.3-common \
  php8.3-mysql php8.3-redis php8.3-mbstring php8.3-xml \
  php8.3-curl php8.3-zip php8.3-bcmath php8.3-intl \
  php8.3-readline php8.3-opcache

# 验证安装
php -v
nginx -v
php-fpm8.3 -v

四、Nginx负载均衡配置

4.1 主负载均衡器配置

# /etc/nginx/nginx.conf
user www-data;
worker_processes auto;
worker_rlimit_nofile 65535;

events {
    worker_connections 4096;
    multi_accept on;
    use epoll;
}

http {
    include       mime.types;
    default_type  application/octet-stream;

    sendfile        on;
    tcp_nopush      on;
    tcp_nodelay     on;
    keepalive_timeout 65;
    client_max_body_size 50M;

    # Gzip压缩
    gzip on;
    gzip_types text/plain text/css application/json application/javascript text/xml;
    gzip_min_length 1000;

    # ThinkPHP应用上游集群
    upstream thinkphp_cluster {
        # 最少连接策略
        least_conn;
        server 192.168.1.110:9000 weight=5 max_fails=3 fail_timeout=30s;
        server 192.168.1.111:9000 weight=5 max_fails=3 fail_timeout=30s;
        keepalive 32;
    }

    # 静态资源上游
    upstream static_cluster {
        server 192.168.1.110:80;
        server 192.168.1.111:80;
    }

    server {
        listen 80;
        server_name example.com www.example.com;

        # 静态文件直接由应用节点处理
        location /static/ {
            proxy_pass http://static_cluster;
            proxy_cache_valid 200 7d;
            expires 7d;
        }

        # PHP请求转发到PHP-FPM集群
        location ~ \.php$ {
            fastcgi_pass thinkphp_cluster;
            fastcgi_index index.php;
            fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
            include fastcgi_params;

            # 超时设置
            fastcgi_connect_timeout 300;
            fastcgi_send_timeout 300;
            fastcgi_read_timeout 300;
        }

        # 伪静态规则（ThinkPHP）
        location / {
            if (!-e $request_filename) {
                rewrite ^(.*)$ /index.php?s=$1 last;
                break;
            }
        }

        # 健康检查
        location /health {
            access_log off;
            return 200 "OK";
        }
    }
}

4.2 Keepalived高可用

在LB-1和LB-2上安装Keepalived：

sudo apt install keepalived -y

主节点配置（LB-1）：

# /etc/keepalived/keepalived.conf
global_defs {
    router_id NGINX_LB
}

vrrp_script chk_nginx {
    script "/usr/local/bin/check_nginx.sh"
    interval 2
    weight -20
    fall 3
    rise 2
}

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass your_secret
    }
    virtual_ipaddress {
        192.168.1.50/24
    }
    track_script {
        chk_nginx
    }
}

备用节点配置（LB-2）：将state改为BACKUP，priority改为90。

健康检查脚本：

#!/bin/bash
# /usr/local/bin/check_nginx.sh
if ! killall -0 nginx; then
    systemctl restart nginx
    sleep 2
    if ! killall -0 nginx; then
        systemctl stop keepalived
    fi
fi

五、ThinkPHP部署与优化

5.1 部署应用代码

# 所有应用节点执行
cd /var/www

# 克隆代码
git clone https://your-repo/thinkphp-app.git app

# 设置权限
chown -R www-data:www-data app
chmod -R 755 app

# 安装Composer依赖
cd app
composer install --no-dev --optimize-autoloader

# 生成运行时目录
mkdir -p runtime
chmod -R 777 runtime

5.2 PHP-FPM优化配置

; /etc/php/8.3/fpm/pool.d/www.conf
[www]
user = www-data
group = www-data
listen = 9000
listen.owner = www-data
listen.group = www-data

; 进程管理
pm = dynamic
pm.max_children = 80
pm.start_servers = 20
pm.min_spare_servers = 10
pm.max_spare_servers = 40
pm.max_requests = 1000
pm.process_idle_timeout = 10s

; 性能优化
php_admin_value[opcache.enable] = 1
php_admin_value[opcache.memory_consumption] = 256
php_admin_value[opcache.max_accelerated_files] = 20000
php_admin_value[opcache.revalidate_freq] = 60
php_admin_value[opcache.validate_timestamps] = 0

; 超时设置
request_terminate_timeout = 300

5.3 ThinkPHP配置优化

// config/database.php - 数据库读写分离
return [
    'default' => 'mysql',
    'connections' => [
        'mysql' => [
            'type' => 'mysql',
            'hostname' => '192.168.1.120',
            'database' => 'thinkphp',
            'username' => 'app_user',
            'password' => 'YourPassword',
            'hostport' => '3306',
            'debug' => false,
            // 主从配置
            'deploy' => 1,
            'rw_separate' => true,
            'master_num' => 1,
            'slave_no' => '',
            'hosts' => [
                // 读从库
                ['hostname' => '192.168.1.121', 'weight' => 1],
            ],
            // 连接池
            'break_reconnect' => true,
            'max_connections' => 50,
        ],
    ],
];

// config/cache.php - Redis缓存
return [
    'default' => 'redis',
    'stores' => [
        'redis' => [
            'type' => 'redis',
            'host' => '192.168.1.130',
            'port' => 6379,
            'password' => '',
            'select' => 0,
            'timeout' => 0,
            'expire' => 3600,
            'persistent' => true,
            'prefix' => 'tp_cache:',
        ],
    ],
];

// config/session.php - Redis会话
return [
    'type' => 'redis',
    'store' => null,
    'host' => '192.168.1.130',
    'port' => 6379,
    'password' => '',
    'select' => 1,
    'expire' => 7200,
    'prefix' => 'tp_sess:',
];

六、数据库高可用

6.1 MariaDB主从配置

主库（DB-M）：

# /etc/mysql/mariadb.conf.d/50-server.cnf
[mysqld]
server-id = 1
log_bin = mysql-bin
binlog_format = ROW
expire_logs_days = 7
max_binlog_size = 100M

# 半同步复制
plugin_load_add = semisync_master.so
rpl_semi_sync_master_enabled = 1
rpl_semi_sync_master_timeout = 1000

从库（DB-S）：

# /etc/mysql/mariadb.conf.d/50-server.cnf
[mysqld]
server-id = 2
relay_log = relay-bin
log_bin = mysql-bin
read_only = 1

plugin_load_add = semisync_slave.so
rpl_semi_sync_slave_enabled = 1

6.2 创建复制用户

-- 在主库执行
CREATE USER 'repl_user'@'%' IDENTIFIED BY 'ReplPassword123!';
GRANT REPLICATION SLAVE ON *.* TO 'repl_user'@'%';
FLUSH PRIVILEGES;

-- 查看主库状态
SHOW MASTER STATUS;

6.3 启动从库同步

-- 在从库执行
CHANGE MASTER TO
    MASTER_HOST='192.168.1.120',
    MASTER_USER='repl_user',
    MASTER_PASSWORD='ReplPassword123!',
    MASTER_LOG_FILE='mysql-bin.000001',
    MASTER_LOG_POS=154;

START SLAVE;
SHOW SLAVE STATUS\G

七、文件同步方案

7.1 使用NFS共享上传文件

存储服务器：

sudo apt install nfs-kernel-server -y

# 创建共享目录
mkdir -p /data/shared/uploads
chown -R nobody:nogroup /data/shared/uploads

# 配置NFS导出
echo "/data/shared/uploads 192.168.1.0/24(rw,sync,no_subtree_check,no_root_squash)" >> /etc/exports
sudo exportfs -a
sudo systemctl restart nfs-kernel-server

应用节点挂载：

sudo apt install nfs-common -y
sudo mkdir -p /var/www/app/public/uploads
echo "192.168.1.120:/data/shared/uploads /var/www/app/public/uploads nfs defaults 0 0" >> /etc/fstab
sudo mount -a

7.2 使用lsyncd实时同步（替代方案）

sudo apt install lsyncd -y

# /etc/lsyncd/lsyncd.conf.lua
settings {
    logfile = "/var/log/lsyncd/lsyncd.log",
    statusFile = "/var/log/lsyncd/lsyncd.status",
    inotifyMode = "CloseWrite",
    maxProcesses = 10,
}

sync {
    default.rsync,
    source = "/var/www/app/public/uploads",
    target = "192.168.1.111:/var/www/app/public/uploads",
    delay = 1,
    rsync = {
        archive = true,
        compress = true,
        perms = true,
    },
}

八、监控与告警

8.1 Prometheus + Grafana监控

监控关键指标：

指标	告警阈值	说明
PHP-FPM进程数	>80% max_children	需要扩容
请求响应时间	>3秒	性能瓶颈
错误率	>1%	应用异常
MariaDB复制延迟	>5秒	数据不一致
磁盘使用率	>85%	需要清理
内存使用率	>90%	OOM风险

8.2 应用健康检查

// app/app/controller/Health.php
namespace app\controller;

class Health
{
    public function index()
    {
        $checks = [
            'database' => $this->checkDatabase(),
            'redis' => $this->checkRedis(),
            'disk' => $this->checkDisk(),
        ];

        $healthy = !in_array(false, $checks);
        return json([
            'status' => $healthy ? 'OK' : 'FAIL',
            'checks' => $checks,
            'timestamp' => date('Y-m-d H:i:s'),
        ], $healthy ? 200 : 503);
    }

    private function checkDatabase(): bool
    {
        try {
            \think\facade\Db::query('SELECT 1');
            return true;
        } catch (\Exception $e) {
            return false;
        }
    }

    private function checkRedis(): bool
    {
        try {
            \think\facade\Cache::set('health_check', 1, 10);
            return (bool)\think\facade\Cache::get('health_check');
        } catch (\Exception $e) {
            return false;
        }
    }

    private function checkDisk(): bool
    {
        $free = disk_free_space('/var/www/app');
        $total = disk_total_space('/var/www/app');
        return ($free / $total) > 0.15;
    }
}

九、自动化部署脚本

#!/bin/bash
# deploy.sh - 一键部署新版本到所有节点
# 用法: ./deploy.sh v1.2.3

VERSION=$1
NODES=("192.168.1.110" "192.168.1.111")
APP_DIR="/var/www/app"
BACKUP_DIR="/backup/deploy"

echo "=== 开始部署版本 $VERSION ==="

for node in "${NODES[@]}"; do
    echo ">>> 部署到 $node ..."

    # 创建备份
    ssh root@$node "cd $APP_DIR && tar -czf $BACKUP_DIR/pre_$VERSION_$(date +%Y%m%d%H%M).tar.gz ."

    # 拉取代码
    ssh root@$node "cd $APP_DIR && git fetch origin && git checkout $VERSION && composer install --no-dev"

    # 清理缓存
    ssh root@$node "cd $APP_DIR && php think clear && php think optimize:route"

    # 重启PHP-FPM
    ssh root@$node "systemctl reload php8.3-fpm"

    echo ">>> $node 部署完成"
done

echo "=== 所有节点部署完成 ==="

十、总结

组件	方案	高可用策略
负载均衡	Nginx + Keepalived	VIP漂移，故障自动转移
应用层	PHP-FPM多节点	水平扩展，滚动更新
数据库	MariaDB主从	半同步复制，自动切换
会话	Redis	单点+持久化，Sentinel可选
文件	NFS / lsyncd	共享存储或实时同步
缓存	Redis	单实例或集群模式

关键原则：
1. 每一层都要有冗余，消除单点故障
2. 监控覆盖所有组件，告警及时响应
3. 自动化部署，减少人为错误
4. 定期演练故障恢复流程

注：本文基于ThinkPHP 8.x、PHP 8.3、Debian 12编写，请根据实际环境调整配置。