Working with Patterns¶
Topics
Patterns in Ansible are how we decide which hosts to manage. This can mean what hosts to communicate with, but in terms of Working With Playbooks it actually means what hosts to apply a particular configuration or IT process to.
We’ll go over how to use the command line in Introduction To Ad-Hoc Commands section, however, basically it looks like this:
ansible <pattern_goes_here> -m <module_name> -a <arguments>
Such as:
ansible webservers -m service -a "name=httpd state=restarted"
A pattern usually refers to a set of groups (which are sets of hosts) – in the above case, machines in the “webservers” group.
Anyway, to use Ansible, you’ll first need to know how to tell Ansible which hosts in your inventory to talk to. This is done by designating particular host names or groups of hosts.
The following patterns are equivalent and target all hosts in the inventory:
all
*
It is also possible to address a specific host or set of hosts by name:
one.example.com
one.example.com:two.example.com
192.0.2.50
192.0.2.*
The following patterns address one or more groups. Groups separated by a colon indicate an “OR” configuration. This means the host may be in either one group or the other:
webservers
webservers:dbservers
You can exclude groups as well, for instance, all machines must be in the group webservers but not in the group phoenix:
webservers:!phoenix
You can also specify the intersection of two groups. This would mean the hosts must be in the group webservers and the host must also be in the group staging:
webservers:&staging
You can do combinations:
webservers:dbservers:&staging:!phoenix
The above configuration means “all machines in the groups ‘webservers’ and ‘dbservers’ are to be managed if they are in the group ‘staging’ also, but the machines are not to be managed if they are in the group ‘phoenix’ … whew!
You can also use variables if you want to pass some group specifiers via the “-e” argument to ansible-playbook, but this is uncommonly used:
webservers:!{{excluded}}:&{{required}}
You also don’t have to manage by strictly defined groups. Individual host names, IPs and groups, can also be referenced using wildcards
*.example.com
*.com
It’s also ok to mix wildcard patterns and groups at the same time:
one*.com:dbservers
You can select a host or subset of hosts from a group by their position. For example, given the following group:
[webservers]
cobweb
webbing
weber
You can refer to hosts within the group by adding a subscript to the group name:
webservers[0] # == cobweb
webservers[-1] # == weber
webservers[0:2] # == webservers[0],webservers[1]
# == cobweb,webbing
webservers[1:] # == webbing,weber
webservers[:3] # == cobweb,webbing,weber
Most people don’t specify patterns as regular expressions, but you can. Just start the pattern with a ‘~’:
~(web|db).*\.example\.com
While we’re jumping a bit ahead, additionally, you can add an exclusion criteria just by supplying the --limit
flag to /usr/bin/ansible or /usr/bin/ansible-playbook:
ansible-playbook site.yml --limit datacenter2
And if you want to read the list of hosts from a file, prefix the file name with ‘@’.:
ansible-playbook site.yml --limit @retry_hosts.txt
Easy enough. See Introduction To Ad-Hoc Commands and then Working With Playbooks for how to apply this knowledge.
Note
You can use ‘,’ instead of ‘:’ as a host list separator. The ‘,’ is preferred specially when dealing with ranges and ipv6.
See also
- Introduction To Ad-Hoc Commands
- Examples of basic commands
- Working With Playbooks
- Learning ansible’s configuration management language
- Mailing List
- Questions? Help? Ideas? Stop by the list on Google Groups
- irc.freenode.net
- #ansible IRC chat channel