capistrano 和 cron 中的 Rake 任务使用错误的 ruby​​ 版本

Question

我需要在部署过程中执行一个 rake 任务。 为此，我在 deploy 命名空间中有一个任务：

desc 'Start Telegram server'
task :start_telegram do
  on roles(:app) do
    puts "Current path: #{release_path}"
    within release_path do
      with rails_env: fetch(:rails_env) do
        execute :nohup, "bundle exec rake bot:start_poller > #{shared_path}/log/telegram.log 2>&1 & echo $! > #{shared_path}/tmp/pids/telegram-server.pid"
      end
    end
  end
end

它是这样执行的： after :stop_telegram, :start_telegram 。 日志文件的输出是：

/var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/spec_set.rb:87:in `block in materialize': Could not find nokogiri-1.10.3 in any of the sources (Bundler::GemNotFound)
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/spec_set.rb:81:in `map!'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/spec_set.rb:81:in `materialize'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/definition.rb:170:in `specs'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/definition.rb:237:in `specs_for'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/definition.rb:226:in `requested_specs'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/runtime.rb:108:in `block in definition_method'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/runtime.rb:20:in `setup'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler.rb:107:in `setup'
        from /var/lib/gems/2.3.0/gems/bundler-2.0.2/lib/bundler/setup.rb:20:in `<top (required)>'
        from /usr/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in `require'
        from /usr/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in `require'

rbenv ruby​​ 版本是 2.5.3， ruby -v给了我ruby 2.5.3p105 (2018-10-18 revision 65156) [x86_64-linux] ，我不明白为什么它会查看/usr/lib/gems/2.3.0 。 当 cron 执行 rake 任务时，也会发生同样的情况。 关于什么可能是错误的任何想法？ 先谢谢了。

编辑：如果我在服务器上手动执行它： RAILS_ENV=production nohup bundle exec rake bot:start_poller > /home/deployer/apps/aws-rails/www/shared/log/telegram.log 2>&1 & echo $! > /home/deployer/apps/aws-rails/www/shared/tmp/pids/telegram-server.pid RAILS_ENV=production nohup bundle exec rake bot:start_poller > /home/deployer/apps/aws-rails/www/shared/log/telegram.log 2>&1 & echo $! > /home/deployer/apps/aws-rails/www/shared/tmp/pids/telegram-server.pid ，一切正常。

Answer 1

如果您希望 Capistrano 使用您的 rvm 版本，那么您需要使用 capistrano/rbenv gem

# Capfile
require 'capistrano/rbenv'


# config/deploy.rb
set :rbenv_type, :user # or :system, depends on your rbenv setup
set :rbenv_ruby, '2.4.2'

欲了解更多详情，请查看以下链接

https://github.com/capistrano/rbenv

要在 crons 中使用 rbenv，请将以下内容添加到您的 bashrc

if which rbenv > /dev/null; then eval "$(rbenv init -)"; fi

https://benscheirman.com/2013/12/using-rbenv-in-cron-jobs/

Answer 2

我找到了解决方案，应该添加这个$HOME/.rbenv/bin/rbenv exec ，所以，最后，可执行命令应该是这样的： execute :nohup, "$HOME/.rbenv/bin/rbenv exec bundle exec rake bot:start_poller > #{shared_path}/log/telegram.log 2>&1 & echo $! > #{shared_path}/tmp/pids/telegram-server.pid"