curl - 通过capistrano V3安装RVM，CURL在终端中抛出垃圾

Question

我正在尝试进行设置，以便在拥有新服务器时，可以使用capistrano部署设置，而不必使用镜像。

我的问题很多，我认为这是我使用Capistrano V3造成的。例如，我的第一个想法是使用rvm-capistrano gem，但出现以下问题。我确实在我的deploy.rb文件和capfile文件中都包含要求“rvm / capistrano”，以确保确定。

cap staging rvm:install_rvm
cap aborted!
Don't know how to build task 'rvm:install_rvm'

因此，我环顾四周，结果发现有一个capistrano-rvm gem专门用于capistrano v3，但它似乎尚未开发，并且存在相同的问题，capistrano无法理解rvm:install_rvm。

所以以后我决定只是尝试在任务本身中安装rvm，

sudo "apt-get -y install curl"
execute "\curl -L https://get.rvm.io | bash"
execute "source ~/.rvm/scripts/rvm"
execute "rvm requirements"
execute "rvm install ruby"
execute "rvm use ruby --default"
execute "rvm rubygems current"
execute "gem install rails"

但我明白了，不知道为什么curl中的字母cu消失了:

INFO [7e26eda7] Running /usr/bin/env rl -L https://get.rvm.io | bash on 192.168.1.126
DEBUG [7e26eda7] Command: rl -L https://get.rvm.io | bash
DEBUG [7e26eda7]        bash: $'\025rl': command not found
DEBUG [7e26eda7]
 INFO [7e26eda7] Finished in 0.006 seconds command successful.
 INFO [14e12c80] Running /usr/bin/env source ~/.rvm/scripts/rvm on 192.168.1.126
DEBUG [14e12c80] Command: source ~/.rvm/scripts/rvm
 INFO [14e12c80] Finished in 0.153 seconds command successful.
 INFO [a501984d] Running /usr/bin/env rvm requirements on 192.168.1.126
DEBUG [a501984d] Command: rvm requirements
cap aborted!
rvm requirements stdout: Nothing written
rvm requirements stderr: Nothing written
/home/hak8or/.rvm/gems/ruby-2.0.0-p247/gems/sshkit-0.0.34/lib/sshkit/command.rb:94:in `exit_status='
/home/hak8or/.rvm/gems/ruby-2.0.0-p247/gems/sshkit-0.0.34/lib/sshkit/backends/netssh.rb:125:in `block (4 levels) in _execute'

所以经过一番谷歌搜索后，我尝试

execute "curl -L get.rvm.io | bash -s stable --auto"

但作为返回，我得到了

 INFO [0e84ae63] Running /usr/bin/env source ~/.rvm/scripts/rvm on 192.168.1.126
DEBUG [0e84ae63] Command: source ~/.rvm/scripts/rvm
 INFO [0e84ae63] Finished in 0.139 seconds command successful.
 INFO [9e5a9f02] Running /usr/bin/env rvm requirements on 192.168.1.126
DEBUG [9e5a9f02] Command: rvm requirements
cap aborted!
rvm requirements stdout: Nothing written
rvm requirements stderr: Nothing written
/home/hak8or/.rvm/gems/ruby-2.0.0-p247/gems/sshkit-0.0.34/lib/sshkit/command.rb:94:in `exit_status='
/home/hak8or/.rvm/gems/ruby-2.0.0-p247/gems/sshkit-0.0.34/lib/sshkit/backends/netssh.rb:125:in `block (4 levels) in _execute'

和curl命令中的 garbage吨。我认为RVM安装不正确。

如果我只有capistrano运行包含所有这些命令的脚本，那么我在终端窗口中也会得到大量垃圾。但是，当我在本地运行它时，这意味着我要进入计算机并运行

\curl -L https://get.rvm.io | bash
rvm requirements

我没问题。我认为这与通过Capistrano专门通过ssh连接运行有关，但没有任何线索。我正在完全更新的ubuntu服务器上的部署用户上运行所有这些操作，并且出于测试目的，我向部署用户授予了完全无密码的sudo。

Answer 1

这是我用来构建jenkins服务器的完整的Install Ruby功能。请记住，我使用Capistrano V2。请享用。

task :install_ruby, roles => :app  do
  run "yum --exclude=*.i386 --exclude=*.i586 install -y gcc-c++ patch readline readline-devel zlib zlib-devel libyaml-devel libffi-devel libxml2-devel libxslt-devel"
  run "bash -s stable < <(curl -k -s https://raw.github.com/wayneeseguin/rvm/master/binscripts/rvm-installer)"
  run "source /etc/profile"  
  run "rvm install 1.9.3 --autolibs=enabled"
  run "rvm use 1.9.3 --default"   
  run "gem install capistrano", { :shell => 'bash'}
  run "gem install railsless-deploy", { :shell => 'bash'}
  run "gem install ntlm-http; true", { :shell => 'bash'}
  run "gem install domain_name", { :shell => 'bash'}
  run "gem install webrobots", { :shell => 'bash'}
  run "gem install mechanize", { :shell => 'bash'}
end