WordPress报502错误问题解决-php-fpm-84.service loaded failed failed LSB: starts php-fpm

问题描述

服务器环境：

• php：8.4
• MySQL：8.0
• Nginx：1.26.2

在访问站点时，一直报502，而两天前还能正常访问。

问题排查

导致502的问题很多，比如站点访问量太大，站点受到了攻击等，但是我看了后台，重启了Nginx，也没解决。现在遇到问题，首选就用AI来诊断。

我先复制了网站的错误日志：

2025/03/07 13:16:26 [crit] 795740#0: *2451 connect() to unix:/tmp/php-cgi-84.sock failed (2: No such file or directory) while connecting to upstream, client: 66.249.66.21, server: www.imqd.cn, request: ＂GET /robots.txt HTTP/1.1＂, upstream: ＂fastcgi://unix:/tmp/php-cgi-84.sock:＂, host: ＂www.imqd.cn＂
2025/03/07 13:16:26 [crit] 795740#0: *2450 connect() to unix:/tmp/php-cgi-84.sock failed (2: No such file or directory) while connecting to upstream, client: 66.249.66.1, server: www.imqd.cn, request: ＂GET /tag/es6 HTTP/1.1＂, upstream: ＂fastcgi://unix:/tmp/php-cgi-84.sock:＂, host: ＂www.imqd.cn＂

丢给ChatGPT，然后它回复的解决方案有点多，我用了下面这一条：

# 将该指令复制，并在站点目录下执行命令行
systemctl list-units --type=service | grep php

输出：

php-fpm-74.service                       loaded active running LSB: starts php-fpm
● php-fpm-84.service                     loaded failed failed  LSB: starts php-fpm

错误如上：

表示我的环境中安装了2个php版本，其中Nginx 试图使用 PHP 8.4 (php-fpm-84)，但它 无法启动，所以 502 Bad Gateway。

问题解决

这个我之前是用PHP8.4来运行WordPress站点的，现在它的服务遇到了问题，导致502问题。

所以很简单的解决方式就是将WordPress的PHP版本切换为7.4即可。

但是为什么会导致8.4 用不了呢？wp运行在PHP8中更安全。

查看PHP的日志:

[14-Feb-2025 04:04:02] NOTICE: child 322153 stopped for tracing
[14-Feb-2025 04:04:02] NOTICE: about to trace 322153
[14-Feb-2025 04:04:02] NOTICE: finished trace of 322153
[15-Feb-2025 17:18:04] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 8 children, there are 1 idle, and 22 total children
[15-Feb-2025 17:18:05] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 16 children, there are 3 idle, and 26 total children
[16-Feb-2025 17:52:51] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 8 children, there are 2 idle, and 27 total children
[16-Feb-2025 17:52:52] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 16 children, there are 3 idle, and 30 total children
[16-Feb-2025 17:52:53] WARNING: [pool www] seems busy (you may need to increase pm.start_servers, or pm.min/max_spare_servers), spawning 32 children, there are 2 idle, and 32 total children
[17-Feb-2025 04:01:49] WARNING: [pool www] child 1221397, script '/www/wwwroot/www.cn/wp-admin/admin-ajax.php' (request: "GET /wp-admin/admin-ajax.php?action=wordfence_doScan&isFork=0&scanMode=standard&cronKey=c1d5082357026a61a2ce6e9feff7f549&signature=31aa3d000f5f2053ad504322212ad58a52a5a0e3032b25317bbdc7ef656b0606") executing too slow (34.708130 sec), logging
[17-Feb-2025 04:01:49] NOTICE: child 1221397 stopped for tracing
[17-Feb-2025 04:01:49] NOTICE: about to trace 1221397
[17-Feb-2025 04:01:49] NOTICE: finished trace of 1221397
[17-Feb-2025 10:57:21] NOTICE: Terminating ...
[17-Feb-2025 10:57:21] NOTICE: exiting, bye-bye!

发现了2个问题：

PHP-FPM 进程池（pool www）资源不足，服务器负载过高

• 日志多次提示 seems busy，说明 PHP 进程太少，导致请求等待时间变长。
• 你可能需要 增加 PHP-FPM 的进程数（pm.start_servers、pm.min/max_spare_servers）。

某些 PHP 脚本执行太慢，导致超时

• 日志中显示 admin-ajax.php 执行时间过长（34.7 秒），可能是 WordPress 插件（如 Wordfence）在运行扫描任务，影响了服务器性能。

所以归根结底还是服务器配置不够，并且因为我开启了Wordfence插件扫描导致的原因。

---

如果你在开发WordPress时遇到了问题，可以在下方与我联系，作为你的WordPress顾问。