nginx 不停的报 could not bind() 的 error log，实际又不影响访问，是神马情况？

V2EX = way to explore

V2EX 是一个关于分享和探索的地方

现在注册

已注册用户请登录

这是一个创建于 2695 天前的主题，其中的信息可能已经有所发展或是发生改变。

一段时间发现某个文件夹怎么那么大，一看才发现是 nginx 的 error.log 已经 6.5G 啦，因为访问都正常，没发现异常情况，就是不断的出现下面的重复 error log

2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:8888 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:99 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:8888 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:99 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:8888 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:99 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:8888 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:99 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:80 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to [::]:443 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:8888 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: bind() to 0.0.0.0:99 failed (98: Address already in use)
2017/07/02 10:43:27 [emerg] 1766#0: still could not bind()

上面的这些端口 80 99 8888 443 都是在正常访问的状态，不知道为什么还要记录这样的 error log?

15 条回复 • 2017-07-08 14:26:51 +08:00

wellsc

2017-07-08 00:20:57 +08:00

sudo lsof -i tcp:80 试试看

fuxkcsdn

2017-07-08 00:23:26 +08:00 via iPhone

是不是启动了多个 nginx ？ ps 找找看

johnlui

2017-07-08 00:30:26 +08:00 via Android

目测是启动了多个 master 进程

ETiV

2017-07-08 03:40:28 +08:00 via iPhone

装 gitlab 了？

cxbig

2017-07-08 04:20:52 +08:00 via iPhone

master 进程有多个 +1

ghostheaven

2017-07-08 06:32:57 +08:00 via Android

0.0.0.0 是绑定本机所有 ip，可能有一个 ip 的 80 和 443 被别的占用了

rogwan

2017-07-08 08:48:32 +08:00

@wellsc

查看了下，是 ipv4 和 ipv6 抢抢同一个端口造成的？
# sudo lsof -i tcp:80
nginx 1306 root 9u IPv4 10290 0t0 TCP *:http (LISTEN)
nginx 1306 root 10u IPv6 10291 0t0 TCP *:http (LISTEN)
nginx 1312 nobody 9u IPv4 10290 0t0 TCP *:http (LISTEN)
nginx 1312 nobody 10u IPv6 10291 0t0 TCP *:http (LISTEN)

rogwan

2017-07-08 08:51:49 +08:00

@ETiV

确实是装了 gitlab，但是我分配了一个专门的 9596 端口，没有出现上面的冲突 log 中呀

kn007

2017-07-08 11:24:48 +08:00

ipv4 和 ipv6 同事监听 80，要以这种方式：

listen 80;
listen [::]:80 ipv6only=on;

rogwan

2017-07-08 11:54:38 +08:00

@kn007 是这样写的，这两个没冲突。

@fuxkcsdn
@johnlui

这个机器上跑了一个 nginx web，还跑了一个 gitlab，看了进程记录显示，gitlab 里面也有一个 nginx 的 master 进程在跑，虽然指定了端口给 gitlab 用，但估计 gitlab 的 nginx 还默认监听了其他端口，造成已经被 nginx web 占用的端口就 bind 不了。

wellsc

2017-07-08 12:01:39 +08:00 via iPad

@rogwan 应该是同时跑了两个人 Nginx 造成的，pwdx pid 看一下分别是那两个 nginx

johnlui

2017-07-08 12:03:38 +08:00

netstat -tulpn

一看便知

iConnect

2017-07-08 12:15:08 +08:00

@johnlui 这个是看在监听状态的端口，楼主说的没有 bind 上报错，那就看不到了

@rogwan 看看这个 https://stackoverflow.com/questions/25393370/gitlab-nginx-problems-nginx-emerg-bind-to-0-0-0-080-failed-98-address

johnlui

2017-07-08 14:01:17 +08:00

@iConnect bind 上报错是因为别的进程已经监听了呀。。。用这个命令可以准确的找到是哪个进程监听了。

ETiV

2017-07-08 14:26:51 +08:00 via iPhone

LZ 貌似已经解决了？