解决uWSGI listen queue of socket full

Wed Mar 13 10:23:18 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:19 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:20 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:21 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:22 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:23 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:24 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:25 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:26 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***
Wed Mar 13 10:23:27 2019 - *** uWSGI listen queue of socket "127.0.0.1:59861" (fd: 3) full !!! (101/100) ***

现象如上图，今天服务器报出这样的问题，根据提示信息可以理解为：uwsgi的队列满了，所以需要改下配置信息

uwsgi:
  http: 0.0.0.0:8080
  chdir: /export/server/
  module: sorce.web.wsgi
  processes: 6
  listen : 100
  daemonize: /tmp/uwsgi.log
  pidfile: /tmp/uwsgi.pid
  master: true
  disable-logging : true
  stats: /tmp/uwsgi.socket
  enable-threads: true
  buffer-size: 32768

listen 字段的意思就是队列的长度，默认是100，所以报上面的错误，改大点就可以了。

如果问题没有解决，查看系统相关设置
修改/etc/sysctl.conf文件,添加或者修改这几个参数值

#对于一个经常处理新连接的高负载 web服务环境来说，默认的 128 太小了
net.core.somaxconn = 262144
​#表示SYN队列的长度，默认为1024，加大队列长度为8192，可以容纳更多等待连接的网络连接数
net.ipv4.tcp_max_syn_backlog = 8192
#网卡设备将请求放入队列的长度
net.core.netdev_max_backlog = 65536

执行sysctl -p重新load参数设置，这样会立即生效，并且以后重新启动机器也会生效