a. Search for a copy of the haproxy cookbook by issuing the following commands.
knife cookbook site search haproxy
knife cookbook site show haproxy
b. Download a copy of the haproxy cookbook.
knife cookbook site vendor -d haproxy
See ( http://wiki.opscode.com/display/chef/Chef+Repository#ChefRepository-cookbooks ) for an explanation of the vendor branch pattern used in this command.
c. Find the metadata.rb and list the supported OS's for this cookbook
d. Update the the default.rb recipe in the ../cookbooks/haproxy/recipes directory
In this step we are going to create modify the haproxy cookbook default recipe to make it data driven using the "search" api. Modify the existing default.rb to include the following code.
package "haproxy" do
action :install
end
template "/etc/default/haproxy" do
source "haproxy-default.erb"
owner "root"
group "root"
mode 0644
end
service "haproxy" do
supports :restart => true, :status => true, :reload => true
action [:enable, :start]
end
pool_members = search(:node, "role:#{node[:haproxy][:pool_role]}") || []
template "/etc/haproxy/haproxy.cfg" do
source "haproxy.cfg.erb"
owner "root"
group "root"
mode 0644
variables :pool_members => pool_members
notifies :restart, resources(:service => "haproxy")
end
Make sure you understand the relationship of the search API call and the variables used in the template. We will see later when we build the role for this example how the attribute "haproxy" sets the pool_role that gets resolved in the erb template.
e. Modify the the erb template for the haproxy cookbook.
In this step we are going to modify the default template erb template to match the updates made in the haproxy recipe. At this point you should know how to find the template file. Modify the existing haproxy.cfg template file to include the following code.
global
log 127.0.0.1 local0
log 127.0.0.1 local1 notice
#log loghost local0 info
maxconn 4096
#debug
#quiet
user haproxy
group haproxy
defaults
log global
mode http
option httplog
option dontlognull
retries 3
redispatch
maxconn 2000
contimeout 5000
clitimeout 50000
srvtimeout 50000
# Set up application listeners here.
listen application 0.0.0.0:80
balance roundrobin
<% @pool_members.each do |member| %>
<% server_ip = member.has_key?("ec2") ? member.ec2.public_ipv4 : member.ipaddress %>
server <%= member.hostname %> <%= server_ip %>:80 weight 1 maxconn 1 check
<% end %>
listen admin 0.0.0.0:22002
mode http
stats uri /
In this example we are using the public IP address of the web servers. If the haproxy server and all of the web servers are on Amazon's cloud it is more efficient to use the private IP addresses. The following code is an example of using the the private IP addresses.
# Set up application listeners here.
listen application 0.0.0.0:80
balance roundrobin
<% @pool_members.each do |member| %>
server <%= member.hostname %> <%= member.ipaddress %>:80 weight 1 maxconn 1 check
<% end %>
f. Upload the updated haproxy
knife cookbook upload haproxy
g. Create a LB role
create a lb role
$EDITOR roles/lb.rb
name "lb"
description "load balancer"
override_attributes(
:haproxy => {:pool_role => "<input the correct value here>"}
)
run_list(
"recipe[haproxy]"
)
In this step you need to input the correct value for the override_attributes to coordinate with the previous code specified in the haproxy recipe and template. Remember the search API in the default.rb recipe will use this value to find all of the launched webservers. If you are not sure about this complete process please take this time to talk to your instructor for clarification.
h. Load the new lb.rb role up to the Chef server
knife role from file lb.rb
i. Show the newly loaded role
knife role show lb
j. Bootstrap a second EC2 instance wit the new loadbalancer instance and also add the new lb role.
k. List your running instances and launch the web page using the following commands.
knife status --run-list
knife status "role:lb" --run-list
curl <the public IP of the lb instance>
Note: The IP address displayed from the page should match the IP address of your webserver instance.
Also display the haproxy admin interface.
curl <the public IP of the lb instance>:22002
__l. Take a deep breath.., and smile... you are having fun right?
m. Rerun the chef-client on the haproxy server to update the load balance pool with the new servers.
n. List all of your running instances and launch the web page using the following commands.
knife ec2 server list
knife status --run-list
knife status "role:webserver" --run-list
knife status "role:lb" --run-list
curl <the public IP of the lb instance>
Note: The IP address displayed from the page should match the IP address of your webserver instance.
Also display the haproxy admin interface.
curl <the public IP of the lb instance>:22002
This output should display the three webserver instances.