Skip to content

Latest commit

 

History

History
305 lines (248 loc) · 11.4 KB

README.md

File metadata and controls

305 lines (248 loc) · 11.4 KB
Raw

chef-ark Build Status

Overview

This cookbook provides ark, a resource for managing software archives. It manages the fetch-unpack-configure-build-install process common to installing software from source, or from binary distributions that are not fully fledged OS packages.

This is a modified verion of Infochimp's awesome install_from cookbook. It has been heavily refactored and extended to meet different use cases.

Given a simple project archive available at a url:

ark 'pig' do
  url 'http://apache.org/pig/pig-0.8.0.tar.gz'
end

The provider will:

  • fetch it to to /var/cache/chef/
  • unpack it to the default path (/usr/local/pig-0.8.0)
  • create a symlink for :home_dir (/usr/local/pig) pointing to path
  • add specified binary commands to the enviroment PATH variable

By default, the ark will not run again if the :path is not empty. Ark provides many actions to accommodate different use cases, such as :dump, :cherry_pick, :put, and :install_with_make.

At this time ark only handles files available from URLs using the remote_file provider. It does handle local files using the file:// protocol.

Requirements

This cookbook requires Chef 11 for the provider, as it uses the use_inline_resources method.

More about use_inline_resources in the Chef documentation.

Should work on common Unix/Linux systems with typical userland utilities like tar, gzip, etc. May require the installation of build tools for compiling from source, but that installation is outside the scope of this cookbook.

Attributes

Customize the attributes to suit site specific conventions and defaults.

  • node['ark']['apache_mirror'] - if the URL is an apache mirror, use the attribute as the default.
  • node['ark']['prefix_root'] - default base location if the prefix_root is not passed into the resource.
  • node['ark']['prefix_bin'] - default binary location if the prefix_bin is not passed into the resource.
  • node['ark']['prefix_home'] - default home location if the prefix_home is not passed into the resource.
  • node['ark']['win_install_dir'] - directory where the files will be put on windows
  • node['ark']['package_dependencies'] - prerequisite system packages that need to be installed to support ark.

Resources/Providers

  • ark - does the extract/build/configure dance

Actions

  • :install: extracts the file and creates a 'friendly' symbolic link to the extracted directory path
  • :configure: configure ahead of the install action
  • :install_with_make: extracts the archive to a path, runs make, and make install. It does not run the configure step at this time
  • :dump: strips all directories from the archive and dumps the contained files into a specified path
  • :cherry_pick: extract a specified file from an archive and places in specified path
  • :put: extract the archive to a specified path, does not create any symbolic links
  • :remove: removes the extracted directory and related symlink #TODO
  • :setup_py: runs the command "python setup.py" in the extracted directory
  • :setup_py_build: runs the command "python setup.py build" in the extracted directory
  • :setup_py_install: runs the comand "python setup.py install" in the extracted directory

:cherry_pick

Extract a specified file from an archive and places in specified path.

Relevant Attribute Parameters for :cherry_pick

  • path: directory to place file in.
  • creates: specific file to cherry-pick.

:dump

Strips all directories from the archive and dumps the contained files into a specified path.

NOTE: This currently only works for zip archives

Attribute Parameters for :dump

  • path: path to dump files to.
  • mode: file mode for app_home, as an integer.
    • Example: 0775
  • creates: if you are appending files to a given directory, ark needs a condition to test whether the file has already been extracted. You can specify with creates, a file whose existence indicates the ark has previously been extracted and does not need to be extracted again.

:put

Extract the archive to a specified path, does not create any symbolic links.

Attribute Parameters for :put

  • path: path to extract to.
    • Default: /usr/local
  • append_env_path: boolean, if true, append the ./bin directory of the extracted directory to the global PATH variable for all users.

Attribute Parameters

  • name: name of the package, defaults to the resource name.
  • url: url for tarball, .tar.gz, .bin (oracle-specific), .war, and .zip currently supported. Also supports special syntax :name:version:apache_mirror: that will auto-magically construct download url from the apache mirrors site.
  • version: software version, defaults to 1.
  • mode: file mode for app_home, is an integer.
  • prefix_root: default prefix_root, for use with :install* actions.
  • prefix_home: default directory prefix for a friendly symlink to the path.
    • Example: /usr/local/maven -> /usr/local/maven-2.2.1
  • prefix_bin: default directory to place a symlink to a binary command.
    • Example: /opt/bin/mvn -> /opt/maven-2.2.1/bin/mvn, where the prefix_bin is /opt/bin
  • path: path to extract the ark to. The :install* actions overwrite any user-provided values for :path.
    • Default: /usr/local/<name>-<version> for the :install, :install_with_make actions
  • home_dir: symbolic link to the path :prefix_root/:name-:version, does not apply to :dump, :put, or :cherry_pick actions.
    • Default: :prefix_root/:name
  • has_binaries: array of binary commands to symlink into /usr/local/bin/, you must specify the relative path.
    • Example: [ 'bin/java', 'bin/javaws' ]
  • append_env_path: boolean, similar to has_binaries but less granular. If true, append the ./bin directory of the extracted directory to. the PATH environment variable for all users, by placing a file in /etc/profile.d/. The commands are symbolically linked into /usr/bin/*. This option provides more granularity than the boolean option.
    • Example: mvn, java, javac, etc.
  • environment: hash of environment variables to pass to invoked shell commands like tar, unzip, configure, and make.
  • strip_components: number of components in path to strip when extracting archive. With default value of 1, ark strips the leading directory from an archive, which is the default for both unzip and tar commands.
  • autoconf_opts: an array of command line options for use with the GNU autoconf script.
    • Example: [ '--include=/opt/local/include', '--force' ]
  • make_opts: an array of command line options for use with make.
    • Example: [ '--warn-undefined-variables', '--load-average=2' ]
  • owner: owner of extracted directory.
    • Default: root

Examples

This example copies ivy.tar.gz to /var/cache/chef/ivy-2.2.0.tar.gz, unpacks its contents to /usr/local/ivy-2.2.0/ -- stripping the leading directory, and symlinks /usr/local/ivy to /usr/local/ivy-2.2.0

 # install Apache Ivy dependency resolution tool
 ark "ivy" do
   url 'http://someurl.example.com/ivy.tar.gz'
   version '2.2.0'
   checksum '89ba5fde0c596db388c3bbd265b63007a9cc3df3a8e6d79a46780c1a39408cb5'
   action :install
 end

This example copies jdk-7u2-linux-x64.tar.gz to /var/cache/chef/jdk-7.2.tar.gz, unpacks its contents to /usr/local/jvm/jdk-7.2/ -- stripping the leading directory, symlinks /usr/local/jvm/default to /usr/local/jvm/jdk-7.2, and adds /usr/local/jvm/jdk-7.2/bin/ to the global PATH for all users. The user 'foobar' is the owner of the /usr/local/jvm/jdk-7.2 directory:

 ark 'jdk' do
   url 'http://download.example.com/jdk-7u2-linux-x64.tar.gz'
   version '7.2'
   path "/usr/local/jvm/"
   home_dir "/usr/local/jvm/default"
   checksum  '89ba5fde0c596db388c3bbd265b63007a9cc3df3a8e6d79a46780c1a39408cb5'
   append_env_path true
   owner 'foobar'
 end

Install Apache Ivy dependency resolution tool in /resource_name in this case /usr/local/ivy, do not symlink, and strip any leading directory if one exists in the tarball:

 ark "ivy" do
    url 'http://someurl.example.com/ivy.tar.gz'
    checksum '89ba5fde0c596db388c3bbd265b63007a9cc3df3a8e6d79a46780c1a39408cb5'
    action :put
 end

Install Apache Ivy dependency resolution tool in /home/foobar/ivy, strip any leading directory if one exists:

 ark "ivy" do
   path "/home/foobar"
   url 'http://someurl.example.com/ivy.tar.gz'
   checksum '89ba5fde0c596db388c3bbd265b63007a9cc3df3a8e6d79a46780c1a39408cb5'
   action :put
 end

Strip all directories and dump files into path specified by the path attribute. You must specify the creates attribute in order to keep the extraction from running every time. The directory path will be created if it doesn't already exist:

 ark "my_jars" do
   url  "http://example.com/bunch_of_jars.zip"
   path "/usr/local/tomcat/lib"
   creates "mysql.jar"
   owner "tomcat"
   action :dump
 end

Extract specific files from a tarball (currently only handles one named file):

 ark 'mysql-connector-java' do
   url 'http://oracle.com/mysql-connector.zip'
   creates 'mysql-connector-java-5.0.8-bin.jar'
   path '/usr/local/tomcat/lib'
   action :cherry_pick
 end

Build and install haproxy and use alternave values for prefix_root, prefix_home, and prefix_bin:

 ark "haproxy" do
   url  "http://haproxy.1wt.eu/download/1.5/src/snapshot/haproxy-ss-20120403.tar.gz"
   version "1.5"
   checksum 'ba0424bf7d23b3a607ee24bbb855bb0ea347d7ffde0bec0cb12a89623cbaf911'
   make_opts [ 'TARGET=linux26' ]
   prefix_root '/opt'
   prefix_home '/opt'
   prefix_bin  '/opt/bin'
   action :install_with_make
 end

You can also pass multiple actions to ark and supply the file extension in case the file extension can not be determined by the URL:

 ark "test_autogen" do
   url 'https://github.com/zeromq/libzmq/tarball/master'
   extension "tar.gz"
   action [ :configure, :install_with_make ]
 end

License and Author

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.