将 Sidekiq 升级到 v6 后，它不会在生产中自动启动

Question

这是 Capistrano 的输出：

02:05 sidekiq:start
      01 sudo service sidekiq start index=1
      01 sidekiq (1) start/running, process 26392
    ✔ 01 deployer@IP 0.721s
      02 sudo service sidekiq start index=2
      02 sidekiq (2) start/running, process 26505
    ✔ 02 deployer@IP 0.728s

登录服务器后ps aux | grep sidekiq，没有进程运行ning。所以我回到本地终端选项卡并运行

cap staging sidekiq:start

新的 Sidekiq 进程会在服务器上显示几秒钟：

ps aux | grep sidekiq
deployer   489 52.0  1.0  76344 40856 ?        Rs   13:13   0:00 /home/deployer/.rvm/rubies/ruby-2.6.3/bin/ruby /home/deployer/.rvm/gems/ruby-2.6.3/bin/bundle exec sidekiq -i 1 -e staging
deployer   695  0.0  0.0  10472   936 pts/0    S+   13:13   0:00 grep --color=auto sidekiq
deployer 32744 42.2  1.4 111100 56188 ?        Rs   13:13   0:02 /home/deployer/apps/app-staging/shared/bundle/ruby/2.6.0/bin/sidekiq -i 2 -e staging

但几秒后，它就会消失

如果我运行来自服务器 RAILS_ENV=staging bundle exec sidekiq - Sidekiq 是运行ning。但是第二次我重新启动 server/deploy 新代码，Sidekiq 进程被杀死。

这是我在 Sidekiq 的佣金任务：

namespace :sidekiq do
  desc "Tells Sidekiq (with signal TSTP) it will be shutting down at some point in the near future."\
  " It will stop accepting new work but continue working on current messages"
  task :quiet do
    on roles(:app) do
      puts capture("pgrep -f 'sidekiq' | xargs kill -TSTP")
    end
  end

  desc "Signals that Sidekiq should shut down within the -t timeout option given at start-up (see config/sidekiq.yml)."\
  " It will stop accepting new work, but continue working on current messages (as with USR1)."\
  " Any workers that do not finish within the timeout are forcefully terminated"\
  " and their messages are pushed back to Redis to be executed again when Sidekiq starts up. "
  task :terminate_gracefully do
    on roles(:app) do
      puts capture("pgrep -f 'sidekiq' | xargs kill -TERM")
    end
  end


  desc "Starts sidekiq workers. Fails if there are already running processes."
  task :start do
    on roles(:app) do
      pids =  capture("pgrep -f 'sidekiq'; true")
      puts "Present sidekiq process pids #{pids}"
      if pids.split("\n").count == 1
        # For each worker service with index ID is called
        execute "sudo service sidekiq start index=1"
        execute "sudo service sidekiq start index=2"
      else
        puts "##------------------------------------------------------------------##"
        puts "Sidekiq was not terminated before start execution. Wait untils it is finished and start it again (or kill the processes)."
        puts "##------------------------------------------------------------------##"
      end
    end
  end

  task :restart do
    invoke "sidekiq:terminate_gracefully"
    invoke "sidekiq:start"
  end
end

为什么 Capistrano 没有自动启动 Sidekiq v6？

Answer 1

Sidekiq 6.0 不再依赖守护进程，查看重大更改： https://github.com/mperham/sidekiq/blob/master/Changes.md#60

BREAKING CHANGE Remove the daemonization, logfile and pidfile arguments to Sidekiq. Use a proper process supervisor (e.g. systemd or foreman) to manage Sidekiq. See the Deployment wiki page for links to more resources.

查看此主题如何使用 sidekiq 6.0 正确设置 systemd： https://github.com/seuros/capistrano-sidekiq/issues/224

或官方维基： https://github.com/mperham/sidekiq/wiki/Deployment#running-your-own-process

Answer 2

在 6.0 sidekiq removed daemonization support 中，许多 init 脚本和 capistrano 示例都使用了它。这样做是为了敦促人们使用适当的流程管理器，如 systemd，它提供了许多稳定性和可管理性优势。

您应该相应地修改您的部署（实际上，许多项目没有注意到此更改，因为旧版本也更适合运行这种方式）。请参阅 official wiki 了解如何设置部署。

将 Sidekiq 升级到 v6 后，它不会在生产中自动启动

After upgrading Sidekiq to v6, it doesn't automatically start in production

ruby

ubuntu

capistrano

ruby-on-rails

sidekiq