从 CLI 更新：
cd /var/discourse
git pull
./launcher rebuild app
一切正常，但我遇到了 502 Bad Gateway 错误。
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
6598e9407e29 local_discourse/app "/sbin/boot" 35 minutes ago Up 22 minutes 0.0.0.0:2045->80/tcp app
./launcher logs app
(105) exiting
ok: run: redis: (pid 47) 17s
ok: run: postgres: (pid 44) 17s
supervisor pid: 114 unicorn pid: 118
config/unicorn_launcher: line 71: kill: (118) - No such process
config/unicorn_launcher: line 15: kill: (118) - No such process
(114) exiting
ok: run: redis: (pid 47) 21s
ok: run: postgres: (pid 44) 21s
supervisor pid: 123 unicorn pid: 127
config/unicorn_launcher: line 71: kill: (127) - No such process
config/unicorn_launcher: line 15: kill: (127) - No such process
(123) exiting
ok: run: redis: (pid 47) 26s
ok: run: postgres: (pid 44) 26s
supervisor pid: 132 unicorn pid: 136
curl localhost:2045
<html>
<head><title>502 Bad Gateway</title></head>
<body>
<center><h1>502 Bad Gateway</h1></center>
<hr><center>nginx</center>
</body>
</html>
Discourse 运行在 LXC 容器中，我的第二个论坛也在 LXC 容器中，更新后运行正常。这真的很奇怪。
有什么建议吗？
更新：
我看到某些 Ruby 进程仍在运行并占用大量 CPU：
/usr/local/bin/ruby /usr/local/bin/bundle exec config/unicorn_launcher -E production -c config/unicorn.conf.rb
更新 2：
通过将 LXC 从非特权容器更改为特权容器解决了问题。