Nagios World Conference 2012 presentation introduces Chef and demo'd the Nagios cookbook.

1. Configuring Nagios with Chef
Bryan McLellan
Technical Program Manager, Open Source
btm@opscode.com / @btmspox

2. Overview
• Who am I?
• Why automation
• Introduction to Chef
• Nagios Demo
• Questions

3. Who am I?
• Chef
Early developer, user, pundit
• 10+ years in Systems
Administration
Computer repair, ISPs, Corporate
IT,
Web operations
• Event Logistics Volunteer
Traffic Control, Parking,
Communications, Networking,
Emergency Management
• Hacker-Operator

4. How did we get here?
Bare Metal Deployment

5. How did we get here?
Bare Metal Deployment
• Purchasing

6. How did we get here?
Bare Metal Deployment
• Purchasing
• Vendor build

7. How did we get here?
Bare Metal Deployment
• Purchasing
• Vendor build
• Delivery

8. How did we get here?
Bare Metal Deployment
• Purchasing
• Vendor build
• Delivery
• Installation

9. How did we get here?
Bare Metal Deployment
• Purchasing
• Vendor build
• Delivery
• Installation
• OS deployment

10. How did we get here?
Bare Metal Deployment
• Purchasing
• Vendor build
• Delivery
• Installation
• OS deployment
• Application deployment

11. How did we get here?
Bare Metal Deployment
• Purchasing
• Vendor build
• Delivery Weeks?
• Installation
• OS deployment
• Application deployment

12. How did we get here?
Cloud or Virtual Deployment
• Purchasing
• Vendor build
• Delivery Nearly immediate
• Installation
• OS deployment
• Application deployment

13. How did we get here?
Cloud or Virtual Deployment
• Purchasing
• Vendor build
• Delivery Nearly immediate
• Installation
• OS deployment
• Application deployment
Must be fast

14. Why automate?
Good Reasons:
• More agility and faster scalability
• Improved infrastructure documentation
• Better disaster recovery

15. Why automate?
Good Reasons:
• More agility and faster scalability
• Improved infrastructure documentation
• Better disaster recovery
Really Good Reasons:
• Spend less time on monotonous tasks
• Spend more time solving interesting
problems

16. Why automate?
Operations is responsible for two things:

17. Why automate?
Operations is responsible for two things:
1. Availability

18. Why automate?
Operations is responsible for two things:
1. Availability
2. Efficiency

19. What is Chef?
• Configuration management language
• Systems integration framework
• API for your infrastructure
http://www.flickr.com/photos/morville/3220961040/

20. Chef Principles

21. Chef Principles
• Idempotent

22. Chef Principles
• Idempotent
• Reasonable

23. Chef Principles
• Idempotent
• Reasonable
• Primitives

24. Chef Principles
• Idempotent
• Reasonable
• Primitives
• Scalable

25. Chef Principles
• Idempotent
• Reasonable
• Primitives
• Scalable
• Hackable

26. Chef Principles
• Idempotent
• Reasonable
• Primitives
• Scalable
• Hackable
• Shareable

27. Chef Basics
Chef manages Nodes

28. Chef Basics
Chef manages Nodes
Nodes have Attributes

29. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients

30. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes

31. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes
Each node has a Run List

32. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes
Each node has a Run List
A Run List is a list of Recipes to run

33. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes
Each node has a Run List
A Run List is a list of Recipes to run
A Role also has a Run List

34. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes
Each node has a Run List
A Run List is a list of Recipes to run
A Role also has a Run List
Roles can also be added to a Node’s Run List

35. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes
Each node has a Run List
A Run List is a list of Recipes to run
A Role also has a Run List
Roles can also be added to a Node’s Run List
Nodes can be in Environments

36. Chef Basics
Chef manages Nodes
Nodes have Attributes
Users and Nodes authenticate as Clients
Cookbooks contain Recipes
Each node has a Run List
A Run List is a list of Recipes to run
A Role also has a Run List
Roles can also be added to a Node’s Run List
Nodes can be in Environments
Data bags are... bags of data.

37. Chef Basics Visualized
client: srv01 client: srv02 client: srv03
node: srv01 node: srv02 node: srv03
run_list: “role[web_server]” run_list: “role[web_server]” run_list: “role[db_server]”
role: web_server role: db_server
run_list: [“recipe[apache2]”, “recipe[php]” ] run_list: [ “recipe[mysql]”, “recipe[nfs]” ]

38. Application Programming Interface
The Meatcloud Manifesto
Give me an API or give me death.
-- Andrew Clay Shafer (@littleidea)

39. Chef Stacks
chef-client knife chef-shell chef-solo
API
Open Source Hosted Chef Private Chef

40. Chef 10 Open Source Architecture
Chef Expander

41. Resources
• A Resource is something you manage
service, package, file, user, execute, git

42. Resources
• A Resource is something you manage
service, package, file, user, execute, git
• Resources have actions
start, install, create, deploy
• Resources can notify of other resources

43. Resources
• A Resource is something you manage
service, package, file, user, execute, git
• Resources have actions
start, install, create, deploy
• Resources can notify of other resources
cookbook_file “/etc/apache2/apache2.conf” do
source “apache2.conf”
owner “root”
group “root”
mode 0644
notifies :restart, “service[apache2]”
end

44. Providers
• A Provider performs the actions specified by the resource

45. Providers
• A Provider performs the actions specified by the resource
• Each Resource can have multiple providers
package: apt, yum, macports...
service: upstart, windows, systemd...

46. Providers
• A Provider performs the actions specified by the resource
• Each Resource can have multiple providers
package: apt, yum, macports...
service: upstart, windows, systemd...
• Each platform (OS) has default Providers that can be
package “sudo” do
provider Chef::Provider::Package::Yum
action :install
end

47. A basic recipe
package “apache2” do
action :install
end

48. A basic recipe
package “apache2” do
action :install
end
service “apache2” do
action :enable
end

49. A basic recipe
package “apache2” do
action :install
end
service “apache2” do
action :enable
end
cookbook_file “/etc/apache2/apache2.conf” do
source “apache2.conf”
owner “root”
group “root”
mode 0644
end

50. A basic recipe
package “apache2” do
action :install
end
service “apache2” do
action :enable
end
cookbook_file “/etc/apache2/apache2.conf” do
source “apache2.conf”
owner “root”
group “root”
mode 0644
end
service “apache2” do
action :start
end

51. A basic recipe
package “apache2” do
action :install
end
service “apache2” do
action :enable
supports [ :restart, :reload, :status ]
end
cookbook_file “/etc/apache2/apache2.conf” do
source “apache2.conf”
owner “root”
group “root”
mode 0644
notifies :restart, “service[apache2]”
end
service “apache2” do
action :start
end

52. Search
Search is a first class citizen
# Find all nodes in production that are tagged ‘group_d’
search(:node, “tags:group_d AND chef_environment:prod”)
# Find the mail server in this environment
search (:node, “role:mail_server AND chef_environment:corp”)
Search and Ruby are powerful allies
# Find all my nodes that run on HP hardware
nodes = search(:node, “dmi_systems_manufacturer:HP”)
# Dynamically create a config in a template
<% nodes.each do |node| -%>
server <%= node[‘hostname’] %>
<% end -%>

53. Nagios Demo
•Download nagios server
cookbooks
•Install nagios server
•Create fake nodes

54. knife cookbook site install nagios
knife cookbook upload -a
cd chef-repo
knife role from file monitoring.json
knife data bag create users
knife data bag from file users btm.json
knife node run list add chef-demo-server role[monitoring]
sudo chef-client
http://chef-demo-server/nagios3/
for n in {1..5} ; do knife node from file fake${n}.json ; done
sudo chef-client
knife data bag create nagios_hostgroups
knife data bag from file nagios_hostgroups hp_systems.json
sudo chef-client

55. Questions?
There is lots more to
learn about Chef at
http://