Now that Local Mode ("Zero") has been introduced to Chef (as of 11.8), which offers search out of the box, this repository is going to get a lot less love. I highly reccomend that you consider migrating from Chef Solo to Chef Zero and drop use of this library. That being said, if you're stuck using Chef Solo, little chef-solo-search will still be there for you.
Here's an article walking through the migration:
https://www.chef.io/blog/2014/06/24/from-solo-to-zero-migrating-to-chef-client-local-mode/
Chef-solo-search is a cookbook library that adds data bag search powers to Chef Solo. Data bag support was added to Chef Solo by Chef 0.10.4. Please see Supported queries for a list of query types which are supported.
Install this cookbook into your Chef repository using your favorite cookbook management tool (Librarian, Berkshelf, knife...).
In Chef 11, you must either add this to the run list of the nodes where it's used or include it as a dependency in the recipes that use it. See [changes in Chef 11][].
Now you have to make sure chef-solo knows about data bags, therefore add
data_bag_path "#{node_work_path}/data_bags"
to the config file of chef-solo (defaults to /etc/chef/solo.rb).
The same for your roles, add
role_path "#{node_work_path}/roles"
To support encrypted data bags, add
encrypted_data_bag_secret "path_to_data_bag_secret"
The search methods supports a basic sub-set of the Lucene query language.
# all items in ':users'
search(:users, "*:*")
search(:users)
search(:users, nil)
# all items from ':users' which have a 'username' attribute
search(:users, "username:*")
search(:users, "username:[* TO *]")
# all items from ':users' which don't have a 'username' attribute
search(:users, "(NOT username:*)")
search(:users, "(NOT username:[* TO *])")
# all items from ':users' with username equals 'speedy'
search(:users, "username:speedy")
# all items from ':users' with username is unequal to 'speedy'
search(:users, "NOT username:speedy")
# all items which 'username'-value begins with 'spe'
search(:users, "username:spe*")
# all items which 'children' attribute contains 'tom'
search(:users, "children:tom")
# all items which have at least one element in 'children' which starts with 't'
search(:users, "children:t*")
search(:users, "married:true")
search(:users, "age:35")
search(:users, "age:42 OR age:22")
search(:users, "married:true AND age:35")
search(:users, "children:tom AND (NOT gender:female)")
search(:users, "children:tom AND (NOT gender:female) AND age:42")
The search methods have support for 'roles', 'nodes' and 'databags'.
You can use the standard role objects in JSON form and put them into your role path
{
"name": "monitoring",
"default_attributes": { },
"override_attributes": { },
"json_class": "Chef::Role",
"description": "This is just a monitoring role, no big deal.",
"run_list": [
],
"chef_type": "role"
}
You can also use ruby formatted roles and put them in your role path. The one proviso being that the filename must match the rolename
name "other"
description "AN Other Role"
run_list []
You can use the standard databag objects in JSON form.
{
"id": "my-ssh",
"hostgroup_name": "all",
"command_line": "$USER1$/check_ssh $HOSTADDRESS$"
}
Nodes are injected through a databag called 'node'. Create a databag called 'node' and put your JSON files there.
You can use the standard node objects in JSON form.
{
"id": "vagrant",
"name": "vagrant-vm",
"chef_environment": "_default",
"json_class": "Chef::Node",
"automatic": {
"hostname": "vagrant.vm",
"os": "centos"
},
"normal": {
},
"chef_type": "node",
"default": {
},
"override": {
},
"run_list": [
"role[monitoring]"
]
}
Running tests is as simple as:
% rake test && kitchen test