Question

运行彪马时，出现此错误：

2019-04-22 19:59:10 -0500: Listen loop error: #<Errno::EBADF: Bad file descriptor>
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:386:in `select'
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:386:in `handle_servers'
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:360:in `block in run'
2019-04-22 19:59:10 -0500: Listen loop error: #<Errno::EBADF: Bad file descriptor>
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:386:in `select'
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:386:in `handle_servers'
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:360:in `block in run'
2019-04-22 19:59:10 -0500: Listen loop error: #<Errno::EBADF: Bad file descriptor>
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:386:in `select'
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:386:in `handle_servers'
/Users/my_account/.rvm/gems/ruby-2.5.5/gems/puma-3.12.1/lib/puma/server.rb:360:in `block in run'
2019-04-22 19:59:10 -0500: Listen loop error: #<Errno::EBADF: Bad file descriptor>

我的puma.rb文件：

workers Integer(ENV['WEB_CONCURRENCY'] || 2)
threads_count = Integer(ENV['MAX_THREADS'] || 5)
threads threads_count, threads_count

rackup      DefaultRackup
port        ENV['PORT']     || 3000
environment ENV['RAILS_ENV'] || 'development'

on_restart do
  puts "Puma restarting..."
end

app_dir = File.expand_path("../..", __FILE__)
log_dir = "#{app_dir}/log"

puts "LOG DIR: #{log_dir}"

if ENV['RAILS_ENV'] == 'production'
  stdout_redirect "#{log_dir}/puma.stdout.log", "#{log_dir}/puma.stderr.log", true
end

Causes problem
on_worker_boot do
  ActiveRecord::Base.establish_connection
end

Causes problem
preload_app!

此问题是由puma.rb配置文件引起的。我正在取消注释每一行以查看它们是否可以工作，但是在puma.rb文件中似乎包含任何代码行都只会导致Errno::EBADF: Bad file descriptor错误。

Answer 1

如果使用capistrano，则需要注释preload_app！尝试使用分阶段重新启动而不是重新启动

我通过对puma.rb（配置文件）进行以下更改来解决了此问题。就我而言，我使用capistrano部署并重新启动puma。使用命令 phased-restart 的两个要求是：worker> = 2且不使用预加载

workers 2
#preload_app!

我的重启命令（分阶段重启）：

pumactl -S /srv/apps/my_app/shared/tmp/pids/puma.state -F /srv/apps/my_app/shared/config/puma.rb phased-restart

Puma侦听循环错误：＃<errno :: ebadf：puma.rb配置中的bad =“” file =“”描述符=“”>

1 个答案: