I am using Capistrano to bootstrap an empty Ubuntu VM.
I have a base recipe that installs base requirements:
def set_default(name, *args, &block)
set(name, *args, &block) unless exists?(name)
end
namespace :deploy do
desc "Install everything onto the server"
task :install do
run "#{sudo} apt-get -y update"
# Common dependencies
run "#{sudo} apt-get -y install curl git-core build-essential python-software-properties"
run "#{sudo} apt-get -y install sudo apt-get -y install build-essential openssl libreadline6 libreadline6-dev curl git-core zlib1g zlib1g-dev libssl-dev libyaml-dev libsqlite3-dev sqlite3 libxml2-dev libxslt-dev autoconf libc6-dev ncurses-dev automake libtool bison subversion"
# Set timezone to UTC
run "#{sudo} bash -c 'echo UTC > /etc/timezone'"
run "#{sudo} cp /usr/share/zoneinfo/UTC /etc/localtime"
run "#{sudo} dpkg-reconfigure -f noninteractive tzdata"
end
end
I also have a recipe for installing rvm. I use the rvm-capistrano gem for this:
set_default(:rvm_ruby_string) { "1.9.3@#{application}" }
set_default(:rvm_install_type, :stable)
set_default(:rvm_type, :user)
require 'rvm/capistrano'
after 'deploy:setup', 'rvm:install_rvm'
after 'deploy:setup', 'rvm:install_ruby'
It appears as if the rvm-capistrano gem is modifying the default shell for every task and appending:
rvm_path=$HOME/.rvm/ $HOME/.rvm/bin/rvm-shell '1.9.3@application'
The repercussion of this is that my base recipe is no longer working since it is runned before installing rvm:
cjoudrey (master) app$ cap deploy:install
* executing `deploy:install'
* executing "sudo -p 'sudo password: ' apt-get -y update"
servers: ["33.33.33.10"]
[33.33.33.10] executing command
[33.33.33.10] rvm_path=$HOME/.rvm/ $HOME/.rvm/bin/rvm-shell '1.9.3@app' -c 'sudo -p '\''sudo password: '\'' apt-get -y update'
** [out :: 33.33.33.10] bash: /home/vagrant/.rvm/bin/rvm-shell: No such file or directory
command finished in 45ms
failed: "rvm_path=$HOME/.rvm/ $HOME/.rvm/bin/rvm-shell '1.9.3@app' -c 'sudo -p '\\''sudo password: '\\'' apt-get -y update'" on 33.33.33.10
cjoudrey (master) app$
require 'bundler/capistrano'
load 'config/recipes/base'
load 'config/recipes/rvm'
load 'config/recipes/nginx'
load 'config/recipes/mysql'
load 'config/recipes/nodejs'
server '33.33.33.10', :web, :app, :db, primary: true
set :user, 'vagrant'
set :application, 'app'
set :deploy_to, "/home/#{user}/apps/#{application}"
set :deploy_via, :remote_cache
set :use_sudo, false
set :scm, 'git'
set :repository, '[email protected]:cjoudrey/test.git'
set :branch, 'master'
default_run_options[:pty] = true
ssh_options[:forward_agent] = true
ssh_options[:keys] = `vagrant ssh-config | grep IdentityFile`.split(' ').last
after 'deploy', 'deploy:cleanup'
Has anyone had this issue before? Is there a way to control when rvm-capistrano modifies the default shell? i.e. Is there a way to have it not modify the default shell for deploy:install?
Should I just make my own rvm recipe instead of using rvm-capistrano?
I had a similar problem, and ended up writing a disable_rvm_shell function in my deploy.rb. I documented this in a blog entry: Capistrano, system wide RVM and creating gemsets as part of deploy:setup.