Stork

Stork is a autoinstall utility, kickstart generation tool and server for CentOS and Redhat systems. It aims to fill the gap in the bare metal systems deployment that many of the other tools for cloud and virtual systems excel at.

Installation using rubygems:

$ gem install stork --pre

Install the latest version from the github:

$ git clone https://github.com/rlyon/stork.git

If you are installing the latest version from github, you have two choices to run stork. First, you can run rake install and build/install the gem, or you can run the executables direcly from the bin directory.

storkctl start restart stop [options]

stork host install [name]
stork host list
stork host localboot [name]
stork host reload
stork host show [name]

The syntax for host is as follows:

host "fqdn" do
  attribute "value"
end

where

• fqdn is the fully qualified domain name of the host.
• attribute is the attributes available for this resource.

• layout - Disk layout containing partition and volume group information. You can supply a string or a block value. If a string is supplied stork will attempt to find the id matching a previously defined layout.
• template - The kickstart template to use when generating the autoinstallation instructions
• pxemac - The mac address of the PXE enabled interface. Used to create the boot configuration files.
• pre_snippet - Scripts that will be run in the pre section before the install begins. Snippets are accessed by the basename of the file they are stored in.
• post_snippet - Scripts that will be run in the post section afer the install has successfully completed. Snippets are accessed by the basename of the file they are stored in.
• interface - Network interface information. Takes only a block value.
• distro - Install distribution information. You can supply a string or a block value. If a string is supplied stork will attempt to find the id matching a previously defined distribution.
• timezone - The IANA zone name (e.g. 'America/Chicago').
• firewall - Initial firewall settings. Block only.
• selinux - String or symbol value representing the three selinux states. The only valid values are: enforcing, permissive, or disabled. Default is enforcing.
• package - Adds a package to the install. Generally not needed as the minimal set of packages that are installed by default will be enough to install the configuration management software.
• repos - Add a new repo to the host.
• stork - Url. Override the stork server location.

Typical hosts will look like:

host "server.example.org" do
  template    "default"
  pxemac      "00:11:22:33:44:55"
  layout      "home"
  distro      "centos"
  repo        "whamcloud-client", baseurl: "http://yum.example.com/eln/x86_64"
  package     "foo"


  interface "eth0" do
    bootproto :static
    ip        "99.99.1.8"
    network   "org"
  end

  interface "eth1" do
    bootproto :static
    ip        "192.168.1.10"
    netmask   "255.255.255.0"
    gateway   "192.168.1.1"
    nameserver "192.168.1.253"
    nameserver "192.168.1.252"
  end

  pre_snippet    "setup"
  post_snippet   "ntp"
  post_snippet   "resolv-conf"
  post_snippet   "notify"
end

Or you define hosts programatically with common ruby techniques:

hosts=[
  [10, "00:11:22:33:44:01"],
  [11, "00:11:22:33:44:02"],
  [12, "00:11:22:33:44:03"],
  [13, "00:11:22:33:44:04"],
  [14, "00:11:22:33:44:05"],
  [15, "00:11:22:33:44:06"],
  [16, "00:11:22:33:44:07"],
  [17, "00:11:22:33:44:08"],
  [18, "00:11:22:33:44:09"],
  [19, "00:11:22:33:44:10"]
]

hosts.each do |octet, mac|
  host "c0#{octet}.example.org" do
    template    "default"
    distro      "centos"
    pxemac      mac
    layout      "home"


    interface "eth0" do
      bootproto :static
      ip        "192.168.10.#{octet}"
      network   "org"
    end
  end
end

The syntax for layout is as follows:

layout "name", part: "physical_volume" do
  attribute "value"
end

where

• name is a unique name that can be used to define global resources that can be referenced from other resources by the defined name.
• part is the physical_volume that the volume group will be placed on.
• attribute is the attributes available for this resource.

• zerombr - Initialize invalid partition tables.
• clearpart - Remove partitions prior to the creation of new partitions.
• partition or part - Partition information (see Partition Resource).
• volume_group or vg - Volume group information (see Volume Group Resource).

The syntax for partition is as follows:

partition "mountpoint" do
  attribute "value"
end

where

• mountpoint is the filesystem path where the partition will be mounted. When a volume group will be associated with the partition, use the pv.n naming scheme to specify that the partition is a physical volume.
• attribute is the attributes available for this resource

• size - Size of the partition in MB.
• type - Set the file system type.
• primary - Force allocation of the partition as a primary partition.
• grow - Grow the partition to the maximum amount.
• recommended - Let the installer determine the recommended size.

See layout for an example of how partition can be used to define disk partitions.

The syntax for volume_group is as follows:

volume_group "name" do
  block_attribute "value" do
    attribute "value"
  end
end

where

• name is the name of the volume group.
• attribute is the attributes available for this resource

• logical_volume - Add a logical volume to the volume group.

The syntax for logical_volume is as follows:

logical_volume "name" do
  attribute "value"
end

where

• name is the name of the logical volume.
• attribute is the attributes available for this resource

• path - Mount point.
• size - Size of the logical volume in MB.
• type - Set the file system type.
• grow - Grow the logical volume to the maximum amount.
• recommended - Let the installer determine the recommended size.

Layouts can be defined seperately from hosts and referenced in hosts by name. A typical layout will look like this:

layout "root_and_home" do
  clearpart
  zerombr
  part "/boot" do
    size 100
    type "ext4"
    primary
  end

  part "swap" do
    recommended
    type "swap"
    primary
  end

  part "pv.01" do
    grow
    primary
  end

  volume_group "vg", part: "pv.01" do
    logical_volume "lv_root" do
      path "/"
      type "ext4"
      size 4096
    end

    logical_volume "lv_home" do
      path "/home"
      type "ext4"
      grow
    end
  end
end

The syntax for distro is as follows:

distro "name" do
  attribute "value"
end

where

• name is a unique name that can be used to define global resources that can be referenced from other resources by the defined name.
• attribute is the attributes available for this resource

• kernel - Name of the kernel. This is the kernel that will be transfered via tftp to the host at install time.
• image - Name of the RAM disk image of the installer.
• url - Install url for network install (only supports http - at least thats the only one I'm testing with)

distro "centos" do
  kernel "vmlinuz"
  image "initrd.img"
  url "http://mirror.example.com/centos"
end

The syntax for network is as follows:

network "name" do
  attribute "value"
end

where

• name is a unique name that can be used to define global resources that can be referenced from other resources by the defined name.
• attribute is one or more of the attributes available for this resource.

• netmask - the netmask of the network.
• gateway - the ip address of the network's gateway.
• nameserver - add a nameserver to the network.
• search_path - add a search_path to the network.

network "org" do
  netmask "255.255.255.0"
  gateway "99.99.1.1"
  nameserver "99.99.1.253"
  nameserver "99.99.1.252"
  search_path "example.org"
end

The syntax for interface is as follows:

interface "device" do
  attribute "value"
end

where

• device is the device that is being configured
• attribute is one or more of the attributes available for this resource.

• onboot - enables the interface on boot.
• ipv4 or noipv4 - enable or disable ipv4 support.
• ipv6 or noipv6 - enable or disable ipv6 support.
• defroute or nodefroute - use or don't use the interface for the default route.
• ethtool - a string value representing the ethtool options for the interface.
• bootproto - a string or symbol representing the boot protocol. Allowed values include :static and :dhcp.

For statically configured interfaces: bootproto :static

• ip - the ip address of the interface.
• netmask - the netmask of the network.
• gateway - the ip address of the network's gateway.
• nameserver - add a nameserver to the network.
• search_path - add a search_path to the network.
• network - set the netmask, gateway, nameservers, and search paths with the values found in the specified network resource.

For dynamically configured interfaces: bootproto :dhcp

• dns or nodns - allow or disallow dhcpd to update resolv.conf.

interface "eth0" do
  onboot
  bootproto :dhcp
end

interface "eth1" do
  bootproto :static
  ip        "99.99.1.8"
  network   "org"
end

interface "eth2" do
  bootproto :static
  ip        "192.168.1.10"
  netmask   "255.255.255.0"
  gateway   "192.168.1.1"
  nameserver "192.168.1.253"
  nameserver "192.168.1.252"
end

The syntax for firewall is as follows:

firewall do
  attribute "value"
end

where

• attribute is one or more of the attributes available for this resource.

• enabled - enable the firewall.
• disable - disable the firewall.
• ssh - allow the ssh protocol through the firewall.
• telnet - allow the telnet protocol through the firewall.
• smtp - allow the smtp protocol through the firewall.
• http - allow the http protocol through the firewall.
• ftp - allow the ftp protocol through the firewall.
• trust - add a device (e.g. eth1, eth2, p1p2, etc) as a trusted device.
• allow - allow a port and using the port:protocol format

firewall do
  enabled
  ssh
  allow "8080:tcp"
  trust "eth1"
end

Templates and snippets use the ERB templating system. When the ERB files are rendered, binding are created to expose ruby the underlying ruby objects.

In kickstart templates, the generated kickstart commands can be accessed:

• url - Outputs the kickstart command assigning the URL for network installs.
• network - Outputs each interface kickstart command for all defined interfaces.
• password - Outputs the 'rootpw' command with a randomized password.
• firewall - Outputs the firewall command.
• timezone - Outputs the timezone command for the installer.
• selinux - Outputs the selinux command.
• layout - Outputs all of the partition, volume groups and logical volume commands.
• bootloader - Outputs the bootloader config information.
• zerombr - Outputs the zerombr command.
• clearpart - Outputs the clearpart command.
• repos - Outputs all additional repo commands.
• packages - Outputs the packages section.
• pre_snippets - Renders and outputs all pre snippets in the %pre section.
• post_snippets - Renders and outputs all post snippets in the %post section.

In addition to the kickstart command generators, the following objects are exposed and can be used:

• host - The complete host object for the current host.

Snippets expose the following objects to the template:

• host - The current host.
• authorized_keys - A string containing all the public keys.
• first_boot_content - A string representation of the json content that will make up the first_boot file.
• nameservers - An array of all unique nameservers.
• search_paths - An array of all unique search_paths.
• stork_server - The IP address or hostname of the stork server.
• stork_port - The port that the stork server is running on.
• stork_bind - The bind IP address of the stork server.

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create new Pull Request

To run the kickstart validation tests on your local system, you will need to install python and the python virtualenv module.
I'm currently using 2.7, and I don't know if it makes a difference.
On Linux you can use your favorite package manager if by some chance your distribution didn't come with it installed. On Mac use homebrew or macports to avoid using the system python which is bound to be a very old version.

$ brew install python
$ pip install virtualenv

Once python and virtualenv has been installed, run

$ rake validator:setup

This will create the directories that you need, set up a virtual environment and get everything ready for the integration tests.

$ rake test