I am trying to launch an Amazon EC2 instance following the instructions of Railscast #347 Rubber and Amazon EC2.
The Rails project is configured as follows. The .ruby-version
file:
ruby-1.9.3
Here is the Gemfile
:
source 'http://rubygems.org'
ruby '1.9.3'
gem 'rails', '~> 3.2.13'
gem 'rb-readline', '~> 0.5.0'
gem 'counter_culture', '~> 0.1.12'
gem 'jquery-rails', '~> 3.0.2'
gem 'therubyracer', '~> 0.11.4', require: 'v8'
gem 'attribute_normalizer', '~> 1.1.0'
group :assets do
gem 'sass-rails', '~> 3.2.5'
gem 'coffee-rails', '~> 3.2.2'
gem 'uglifier', '~> 2.1.1'
end
group :development do
gem 'debugger', '~> 1.6.0'
end
group :development, :test do
gem 'sqlite3', '~> 1.3.7'
end
group :production do
gem 'pg', '~> 0.15.1'
end
gem 'rubber', '~> 2.5.2'
gem 'open4'
gem 'gelf'
gem 'graylog2_exceptions', :git => 'git://github.com/wr0ngway/graylog2_exceptions.git'
gem 'graylog2-resque'
I run the following rubber command:
$ rubber vulcanize complete_passenger_postgresql
Besides other files the rubber-ruby.yml
is generated as follows (comments removed):
# config/rubber/rubber-ruby.yml
packages: [build-essential, git-core, subversion, curl, autoconf, bison, ruby, zlib1g-dev, libssl-dev, libreadline6-dev, libxml2-dev, libyaml-dev]
ruby_build_version: 20130501
ruby_version: 1.9.3
ruby_path: "/usr/local/rubies/#{ruby_version}"
After chatting with nirvdrum on IRC I already change the ruby_version: 1.9.3
to ruby_version: 1.9.3-p448
and gave it a try. Same error though.
I sticked to the tutorial and only configured the mandatory settings:
# config/rubber/rubber.yml
app_name: ...
domain: ...
access_key: ...
secret_access_key: ...
account: ...
key_name: ...
Following the tutorial I test the setup running:
$ cap rubber:create_staging
The script runs for a while installing and restarting instances but after a while it stops:
...
** [out :: production.foobar.com] Identified as a _site plugin, moving to _site structure ...
** [out :: production.foobar.com] Installed head
command finished in 5005ms
* 2013-07-11 01:02:20 executing `rubber:graylog:server:install'
servers: ["production.foobar.com"]
** sftp upload #<StringIO:0x007f2d89221718> -> /tmp/install_graylog_server
[production.foobar.com] /tmp/install_graylog_server
[production.foobar.com] done
* sftp upload complete
* executing "sudo -p 'sudo password: ' bash -l /tmp/install_graylog_server"
servers: ["production.foobar.com"]
[production.foobar.com] executing command
command finished in 2320ms
* 2013-07-11 01:02:23 executing `rubber:graylog:web:install'
servers: ["production.foobar.com"]
** sftp upload #<StringIO:0x007f2d817c2a58> -> /tmp/install_graylog_web
[production.foobar.com] /tmp/install_graylog_web
[production.foobar.com] done
* sftp upload complete
* executing "sudo -p 'sudo password: ' bash -l /tmp/install_graylog_web"
servers: ["production.foobar.com"]
[production.foobar.com] executing command
command finished in 945ms
* 2013-07-11 01:02:24 executing `rubber:apache:install'
servers: ["production.foobar.com"]
** sftp upload #<StringIO:0x007f2d81911800> -> /tmp/install_apache
[production.foobar.com] /tmp/install_apache
[production.foobar.com] done
* sftp upload complete
* executing "sudo -p 'sudo password: ' bash -l /tmp/install_apache"
servers: ["production.foobar.com"]
[production.foobar.com] executing command
** [out :: production.foobar.com] Site default disabled.
** [out :: production.foobar.com] To activate the new configuration, you need to run:
** [out :: production.foobar.com] service apache2 reload
command finished in 636ms
* 2013-07-11 01:02:25 executing `rubber:setup_volumes'
* 2013-07-11 01:02:25 executing `rubber:setup_gem_sources'
triggering before callbacks for `rubber:setup_gem_sources'
* 2013-07-11 01:02:25 executing `rubber:base:install_ruby_build'
servers: ["production.foobar.com"]
** sftp upload #<StringIO:0x007f2d81a7f750> -> /tmp/install_ruby_build
[production.foobar.com] /tmp/install_ruby_build
[production.foobar.com] done
* sftp upload complete
* executing "sudo -p 'sudo password: ' bash -l /tmp/install_ruby_build"
servers: ["production.foobar.com"]
[production.foobar.com] executing command
** [out :: production.foobar.com] Installing ruby-build v20130501
** [out :: production.foobar.com] Installed ruby-build at /usr/local
command finished in 1475ms
triggering after callbacks for `rubber:base:install_ruby_build'
* 2013-07-11 01:02:27 executing `rubber:base:install_ruby'
servers: ["production.foobar.com"]
** sftp upload #<StringIO:0x007f2d81aca638> -> /tmp/install_ruby
[production.foobar.com] /tmp/install_ruby
[production.foobar.com] done
* sftp upload complete
* executing "sudo -p 'sudo password: ' bash -l /tmp/install_ruby"
servers: ["production.foobar.com"]
[production.foobar.com] executing command
** [out :: production.foobar.com] Compiling and installing ruby 1.9.3. This may take a while ...
command finished in 1603ms
triggering after callbacks for `rubber:base:install_ruby'
* 2013-07-11 01:02:29 executing `rubber:install_gems'
servers: ["production.foobar.com"]
** sftp upload #<StringIO:0x000000033d5920> -> /tmp/gem_helper
[production.foobar.com] /tmp/gem_helper
[production.foobar.com] done
* sftp upload complete
* executing "sudo -p 'sudo password: ' bash -l -c 'ruby /tmp/gem_helper install $CAPISTRANO:VAR$'"
servers: ["production.foobar.com"]
[production.foobar.com] executing command
** ruby: no such file to load -- ubygems (LoadError)
command finished in 851ms
failed: "/bin/bash -l -c 'sudo -p '\\''sudo password: '\\'' bash -l -c '\\'' \
ruby /tmp/gem_helper install open4 aws-s3 bundler rubber:2.5.2 \
gelf graylog2_exceptions graylog2-resque pg fastthread rack \
passenger:3.0.19 puma'\\'''" on production.foobar.com
Checking for a Ruby installation in /usr/local/rubies
via ssh on the server shows that it has not been installed.
This was posted in the freenode IRC #rubberec2 as a comment on my problem:
However, I cannot figure out how to apply this onto the Rubber configuration.
As stated in the question I generated the configuration files for Rubber including
rubber-ruby.yml
.I repeated the whole process with the same version of Rubber (I just fixed the version to
2.5.2
since there is an update meanwhile). This time I noticed that theruby_version
is set to1.9.3-p392
as you can see here. It seems as if I accidently modified this setting?!Reverting this change the deployment actually works as expected.
As mentioned, in the meantime Rubber 2.5.3 was published which contains updated settings:
I also tried the new version for those who are curious. It works as well.
Big thanks to Kevin Menard and Matt Conway who persevered answering my questions in the IRC channel of the project.