1

I'm working on spinning up a cluster of nodes. In myrepo/cookbooks/mycookbook/.kitchen.yml I have

driver:
  name: vagrant

provisioner:
  name: chef_zero
  roles_path: '../../roles'
  environments_path: '../../environments'
  data_bags_path: '../../data_bags'

platforms:
  - name: ubuntu-12.04
    suites:
      - name: node01
        driver:
          vm_hostname: "node01.localhost"
        run_list:
          - role[genericnode]
        provisioner:
            client_rb:
                environment: development

In myrepo/cookbooks/mycookbook/recipes/default.rb I have

nodes = []
search(:node, 'role:genericnode').each do |node|
    nodes.push("http://#{node['ipaddress']}:8080")
end

node.default['mysetting'] = nodes.join(',')

In my development environment, though, search returns an empty list. Do I need more than the .kitchen.yml settings above to be able to access the ip addresses of the nodes in my network with the role genericnode?

Additional information:

Here's the role:

{
  "name": "genericnode",
  "description": "Generic Node",
  "json_class": "Chef::Role",
  "default_attributes": {

  },
  "override_attributes": {

  },
  "chef_type": "role",
  "run_list": [
    "recipe[apt::default]", "recipe[mycookbook::default]"
  ],
  "env_run_lists": {

  }
}
4

1 に答える 1

1

chef_zero で検索を機能させるには、ノード オブジェクトのフォルダーを JSON として提供し、検索対象のデータにする必要があります。サーバー上にない限り、検索は現在のノードを見つけられません（そのノード/フォルダーに追加しない限り、そうではありません.

これに対する通常の修正は、検索とローカル値のチェックの両方を使用することです。

ips = []
search(:node, 'roles:genericnode').each do |n|
    ips << n['ipaddress']
end
if node['roles'].include?'(genericnode')
  ips |= [node['ipaddress']]
end

node.default['mysetting'] = ips.map{|ip| "http://#{ip}:8080"}.join(',')
于 2015-01-05T23:02:03.730 に答える