Support #691
Updated by Daniel Curtis about 9 years ago
{{>toc}}
This is a simple guide to setup a Puppet Master server along with Puppet Dashboard and Icinga2 with PostgreSQL as the database backend on FreeBSD 9.2.
This will set up 2 web applications that can be viewed in any modern web browser at:
* *Default blank landing page* - http://puppet.example.com
* *Puppet Dashboard* - http://puppet.example.com:3000
* *Icinga* - http://icinga.example.com:8000
* *Puppet Master* - https://puppet.example.com:8140
h1. Prepare the Server
* Once the server baseline has been installed and root access has been obtained make sure the server is up to date:
<pre>
pkg update && upgrade
portsnap fetch extract
</pre>
* Portmaster will be useful for upgrading packages:
<pre>
pkg install portmaster
pkg2ng
</pre>
---
h1. Install PostgreSQL 9.4
* Install PostgreSQL:
<pre>
pkg install postgresql94-{server,client}
</pre>
* Enable PostgreSQL at boot:
<pre>
echo 'postgresql_enable="YES"' >> /etc/rc.conf
</pre>
* Initialize the database:
<pre>
service postgresql initdb
</pre>
* Start PostgreSQL:
<pre>
service postgresql start
</pre>
* Edit the postgres config file:
<pre>
vi /usr/local/pgsql/data/postgresql.conf
</pre>
#* And modify the following:
<pre>
listen_addresses = '*'
</pre>
* Edit the pg_hba config file:
<pre>
vi /usr/local/etc/pgsql/data/pg_hba.conf
</pre>
#* And modify the following:
<pre>
# TYPE DATABASE USER CIDR-ADDRESS METHOD only
# Local connections
local all all trust
# IPv4 local connections:
host all all 127.0.0.1/32 trust
# IPv6 local connections:
host all all ::1/128 trust
# IPv4 connections:
host all all 192.168.10.0/24 md5
# IPv6 local connections:
host all all 1234::abcd/64 md5
</pre>
* Restart postgresql:
<pre>
service postgresql restart
</pre>
h2. Create a new user and database
* Switch to the pgsql user and enter into the psql prompt:
<pre>
su pgsql
psql -d template1
</pre>
#* Create the *puppetmaster* user and database:
<pre>
CREATE USER puppetmasteruser WITH PASSWORD 'SuperSecretPuppetmasterPassword';
CREATE DATABASE puppetmasterdb OWNER puppetmasteruser;
GRANT ALL PRIVILEGES ON DATABASE "puppetmasterdb" to puppetmasteruser;
</pre>
#* Create the *dashboard* user and database:
<pre>
CREATE USER dashboarduser WITH PASSWORD 'SuperSecretDashboardPassword';
CREATE DATABASE dashboarddb OWNER dashboarduser;
GRANT ALL PRIVILEGES ON DATABASE "dashboarddb" to dashboarduser;
</pre>
#* Create the *icinga* user and database:
<pre>
CREATE USER icingauser WITH PASSWORD 'SuperSecretIcingaPassword';
CREATE DATABASE icingadb OWNER icingauser;
GRANT ALL PRIVILEGES ON DATABASE "icingadb" to icingauser;
</pre>
* Exit from the postgres user
<pre>
\q
exit
</pre>
* Test the connection on a remote host:
<pre>
psql -h pg.example.com -U puppetmasteruser -W puppetmasterdb
</pre>
---
h1. Install Puppet Master
* Install the puppet package
<pre>
pkg install puppet rubygem-rake rubygem-bundler rubygem-activerecord rubygem-sqlite3 rubygem-pg libxslt git node portupgrade
</pre>
* Create a puppet config from the package sample:
<pre>
cp /usr/local/etc/puppet/puppet.conf-dist /usr/local/etc/puppet/puppet.conf
</pre>
* Configure your @[puppetmasterd]@ section to reflect these settings:
<pre>
vi /usr/local/etc/puppet/puppet.conf
</pre>
#* And add the following to the bottom of the file:
<pre>
[main]
logdir=/var/log/puppet
vardir=/var/puppet
ssldir=/var/puppet/ssl
rundir=/var/run/puppet
factpath=$vardir/lib/facter
server=puppetmaster.example.com
report=true
pluginsync=true
environment = production
confdir = /usr/local/etc/puppet
[agent]
report = true
show_diff = true
environment = production
[master]
storeconfigs = true
dbadapter = postgresql
dbuser = puppetmasteruser
dbpassword = SuperSecretPuppetmasterPassword
dbserver = localhost
dbname = puppetmasterdb
</pre>
* Create folders for node and class definition files:
<pre>
mkdir /usr/local/etc/puppet/manifests/nodes
mkdir /usr/local/etc/puppet/manifests/classes
</pre>
* Then create a default site.pp file:
<pre>
import "nodes/*.pp"
vi /usr/local/etc/puppet/manifests/site.pp
</pre>
#* And add the following:
<pre>
node default {
notify { "Hey ! It works !": }
}
# The filebucket option allows for file backups to the server
filebucket { main: server => 'puppetmaster.altservice.com', path=> false, }
# Set global defaults - including backing up all files to the main filebucket and adds a global path
File { backup => main }
Exec { path => "/usr/bin:/usr/sbin/:/bin:/sbin" }
</pre>
* Create a puppet node config for *puppetmaster.example.com*:
<pre>
vi /usr/local/etc/puppet/manifests/nodes/puppetmaster.example.com.pp
</pre>
#* And Add the following:
<pre>
node 'puppetmaster.example.com' {
notify { "Hey ! It works !": }
}
</pre>
* Create a puppet node config for *client.example.com*:
<pre>
vi /usr/local/etc/puppet/manifests/nodes/client.example.com.pp
</pre>
#* And Add the following:
<pre>
node 'client.example.com' {
notify { "Hey ! It works !": }
}
</pre>
* Enable puppet in @/etc/rc.conf@:
<pre>
echo 'puppet_enable="YES"' >> /etc/rc.conf
</pre>
* At this point, Puppet needs to be started so that all its SSL keys can be generated. This gives the chance to test that Puppet does work before anything else gets stacked on as well as ensures the SSL keys referenced by Nginx's config file are generated and in place before that step.
<pre>
echo 'puppetmaster_enable="YES"' >> /etc/rc.conf
service puppetmaster start
</pre>
* On *puppetmaster.example.com* - start Puppet on the client system
<pre>
puppet agent -vt
</pre>
h3. Add Pkgng Module
* Install the pkgng module to add the pkgng package provider:
<pre>
puppet module install zleslie-pkgng
</pre>
#* Now packages can be installed using pkgng as the package manager:
<pre>
package { 'puppet':
ensure => installed,
provider => pkgng,
}
</pre>
h3. Enable Puppet File Server
* Make the path for the puppet files to be served from:
<pre>
mkdir /usr/local/etc/puppet/files
</pre>
* Create the puppet fileserver configuration:
<pre>
vi /usr/local/etc/puppet/fileserver.conf
</pre>
#* And put the name of your mount point, the path, and an allow * directive.:
<pre>
[files]
path /usr/local/etc/puppet/files
allow *
</pre>
Next, edit the puppet authorization configuration:
<pre>
vi /usr/local/etc/puppet/auth.conf
</pre>
#* Use a regular expression path to match both the file_metadata and file_content endpoints followed by the name of your custom mount point. Then, use any combination of allow and allow_ip directives to control access. Add the following before the @path /@ statement:
<pre>
path ~ ^/file_(metadata|content)/files/
auth yes
allow /^(.+\.)?example.com$/
allow_ip 192.168.100.0/24
</pre>
h2. Install Puppet on Remote Node
* Install puppet:
<pre>
pkg install puppet portupgrade
</pre>
* Create a basic puppet config:
<pre>
vi /usr/local/etc/puppet/puppet.conf
</pre>
#* And add the following:
<pre>
[main]
logdir=/var/log/puppet
vardir=/var/puppet
ssldir=/var/puppet/ssl
rundir=/var/run/puppet
factpath=$vardir/lib/facter
configtimeout=600
runinterval=28800
report=true
pluginsync=true
server=puppetmaster.example.com
</pre>
* Generate a puppet private key and send the CSR to the puppetmaster node for signing:
<pre>
puppet agent -vt --waitforcert 30
</pre>
h3. Sign the Puppet CSR
* On the *puppetmaster.example.com* node run the following to sign the certi
<pre>
puppet cert sign client.example.com
</pre>
---
h1. Install Puppet Dashboard
Now its time to install Puppet Dashboard, a web frontend to display puppet reports.
* Install puppet-dashboard from git
<pre>
cd /usr/local/www
git clone git://github.com/sodabrew/puppet-dashboard.git
</pre>
* Manually create the 'puppet-dashboard' user and group:
<pre>
pw groupadd -n puppet-dashboard -g 800
pw useradd -n puppet-dashboard -c "Puppet Dashboard,,," -u 800 -g puppet-dashboard -s /usr/sbin/nologin
</pre>
* Then provide the required permissions to @/usr/local/www/puppet-dashboard@
<pre>
chown -R puppet-dashboard:puppet-dashboard /usr/local/www/puppet-dashboard
</pre>
h2. Configure Puppet Dashboard
* Copy the example database YAML file and update with database information:
<pre>
cd /usr/local/www/puppet-dashboard/config
cp database.yml.example database.yml
</pre>
* Then edit the database.yml with the corrected information; make sure to replace the +password+ and +host+:
<pre>
vi database.yml
</pre>
#* and modify the following parameter accordingly:
<pre>
production:
database: dashboarddb
username: dashboarduser
password: SuperSecretDashboardPassword
encoding: utf8
adapter: postgresql
host: localhost
</pre>
* Change the ownership and harden the permissions:
<pre>
chown puppet-dashboard:puppet-dashboard database.yml
chmod 660 database.yml
</pre>
* Copy the example settings YAML file, no changes needed:
<pre>
cd /usr/local/www/puppet-dashboard/config
cp settings.yml.example settings.yml
chown puppet-dashboard:puppet-dashboard settings.yml
chmod 660 settings.yml
</pre>
* Fix shebang line in External Node Classifier Script.
<pre>
sed -i '' -e 's/#! \/usr\/bin\/ruby/#!\/usr\/local\/bin\/ruby/' /usr/local/www/puppet-dashboard/bin/external_node
</pre>
* Install gems required via bundler:
<pre>
cd /usr/local/www/puppet-dashboard
bundle install --path vendor/bundle
</pre>
* Generate secret_token. Cleanup any errors and the default token after generating the new one.
<pre>
echo "secret_token: `bundle exec rake secret`" >> config/settings.yml
vi config/settings.yml
</pre>
* At this point the database was already installed with some blank tables. We need to run rake to finish the process with the database structure needed:
<pre>
cd /usr/local/www/puppet-dashboard
env RAILS_ENV=production bundle exec rake db:setup
</pre>
* Before going into a production environment, Dashboard 2.0 must precompile assets for production:
<pre>
env RAILS_ENV=production bundle exec rake assets:precompile
</pre>
* Chown any files created up until now to the right owner:
<pre>
chown -R puppet-dashboard:puppet-dashboard /usr/local/www/puppet-dashboard
</pre>
* Run Dashboard using Ruby's built-in WEBrick server to validate functionality. It will be available at http://puppet.example.com:3000
<pre>
cd /usr/local/www/puppet-dashboard
su -m puppet-dashboard -c 'env RAILS_ENV=production bundle exec rails server'
</pre>
All agent nodes have to be configured to submit reports to the master. The master has to be configured to send reports to Dashboard. If you already have a working Puppet installation you can configure it to distribute the updated puppet.conf to your hosts.
* Edit the *@puppet.conf@* on the Puppetmaster node:
<pre>
vi /usr/local/etc/puppet/puppet.conf
</pre>
#* And add the following to the [master] section:
<pre>
[master]
storeconfigs = true
dbadapter = postgresql
dbuser = puppetmasteruser
dbpassword = SuperSecretPuppetmasterPassword
dbserver = localhost
dbname = puppetmasterdb
reports = store, http
reporturl = http://puppet.example.com:3000/reports/upload
node_terminus = exec
external_nodes = /usr/bin/env PUPPET_DASHBOARD_URL=http://puppet.example.com:3000 /usr/local/www/puppet-dashboard/bin/external_node
</pre>
* Testing Puppet's Connection to Dashboard. A new background task should show in the Dashboard UI at http://puppet.example.com:3000
<pre>
puppet agent --test
</pre>
* Dashboard ships a worker process manager under script/delayed_job. It can manually start delayed jobs via the following command:
<pre>
su -m puppet-dashboard -c 'env RAILS_ENV=production bundle exec script/delayed_job -p dashboard -n 2 -m start'
</pre>
h2. Delayed Job Worker Init Script
However, rather than manually triggering background workers, this rc script will accomplish the same thing and ensure the background jobs get started on the next reboot.
* Create puppet dashboard FreeBSD init script:
<pre>
vi /usr/local/etc/rc.d/dashboard_workers
</pre>
#* and add the following
<pre>
#!/bin/sh
# PROVIDE: dashboard_workers
# REQUIRE: LOGIN
# KEYWORD: shutdown
# By default dashboard_workers uses flags '-n 1' for 1 worker. This should be
# adjusted to the number of CPU cores.
dashboard_workers_enable=${dashboard_workers_enable:-"NO"}
dashboard_workers_flags=${dashboard_workers_flags:-"-n 1"}
# The default rails environment is set to production
dashboard_workers_env=${dashboard_workers_env:-"/usr/bin/env PATH=${PATH}:/usr/local/bin RAILS_ENV=production"}
# The default user is set to puppet-dashboard and install location is set to
# /usr/local/share/puppet-dashboard.
dashboard_workers_user=${dashboard_workers_user:-"puppet-dashboard"}
dashboard_workers_chdir=${dashboard_workers_chdir:-"/usr/local/www/puppet-dashboard"}
. /etc/rc.subr
name="dashboard_workers"
rcvar="dashboard_workers_enable"
load_rc_config $name
extra_commands="reload run zap status"
# All commands call the same function and strip the fast|one|quiet prefix
# to deliver to the bundler.
reload_cmd="f_dashboard_workers reload"
restart_cmd="f_dashboard_workers restart"
run_cmd="f_dashboard_workers run"
start_cmd="f_dashboard_workers start"
status_cmd="f_dashboard_workers status"
stop_cmd="f_dashboard_workers stop"
zap_cmd="f_dashboard_workers zap"
# Use the function's ARVG $1 as the bundler program's '-m' flag
f_dashboard_workers() {
cd $dashboard_workers_chdir && \
su -m "$dashboard_workers_user" \
-c "${dashboard_workers_env} bundle exec script/delayed_job ${rc_flags} -p dashboard -m $1" || \
echo "Failed to $1 dashboard_workers"
}
run_rc_command "$1"
</pre>
#* And make it executable:
<pre>
chmod +x /usr/local/etc/rc.d/dashboard_workers
</pre>
* With that in place, we need to override the defaults and enable the script along with setting '-n 4' workers to match the number of processor cores and ensure it's ready for a production workload.
<pre>
echo 'dashboard_workers_enable="YES"' >> /etc/rc.conf
echo 'dashboard_workers_flags="-n 4"' >> /etc/rc.conf
service dashboard_workers start
</pre>
---
h1. Puppet Dashboard with Nginx and Passenger
* Install the Passenger gem:
<pre>
portmaster www/rubygem-passenger
</pre>
*NOTE*: Make sure to enable +(*)NGINX and [X]SYMLINK+ when running @make config@
* Install Nginx with Passenger
<pre>
portmaster www/nginx
</pre>
*NOTE*: Make sure to enable +[X]PASSENGER+ during the nginx configuration.
* Create a configuration directory to make managing individual server blocks easier
<pre>
mkdir /usr/local/etc/nginx/conf.d
</pre>
* Configuring Nginx with Passenger, edit the *main nginx configuration file*:
<pre>
vi /usr/local/etc/nginx/nginx.conf
</pre>
#* And add/modify the following
<pre>
user www www;
worker_processes 4;
error_log /var/log/nginx_error.log notice;
pid /var/run/nginx.pid;
events {
worker_connections 1024;
}
http {
passenger_root /usr/local/lib/ruby/gems/2.1/gems/passenger;
passenger_ruby /usr/local/bin/ruby;
passenger_max_pool_size 15;
passenger_pool_idle_time 300;
include mime.types;
default_type application/octet-stream;
sendfile on;
tcp_nopush on;
keepalive_timeout 65;
tcp_nodelay on;
# Load config files from the /etc/nginx/conf.d directory
include /usr/local/etc/nginx/conf.d/*.conf;
}
</pre>
* And create a *puppet dashboard server block*:
<pre>
vi /usr/local/etc/nginx/conf.d/puppet-dashboard.conf
</pre>
#* And add the following:
<pre>
server {
listen 3000;
server_name puppet.example.com;
passenger_enabled on;
passenger_user puppet-dashboard;
passenger_group puppet-dashboard;
access_log /var/log/nginx_dashboard_access.log;
root /usr/local/www/puppet-dashboard/public;
}
</pre>
* Enable a daily log file rotation via @newsyslog.conf@:
<pre>
printf "/var/log/nginx/*.log\t\t\t644 7\t * @T00 JG /var/run/nginx.pid 30\n" >> /etc/newsyslog.conf
</pre>
* Enable nginx service and start it. At this point basic functionality is online:
<pre>
echo 'nginx_enable="YES"' >> /etc/rc.conf
service nginx start
</pre>
* Edit the *@auth.conf@* file on the puppetmaster
<pre>
vi /usr/local/etc/puppet/auth.conf
</pre>
#* Add the following:
<pre>
path /facts
auth yes
method find, search
allow dashboard
</pre>
* Edit the *@site.pp@* on the Puppet master:
<pre>
vi /usr/local/etc/puppet/manifests/site.pp
</pre>
#* Add the following:
<pre>
filebucket { server => "{ main: server => 'puppetmaster.example.com', path=> false, }",
path => false,
}
</pre>
#* In either site.pp, in an individual init.pp, or in a specific manifest.
<pre>
File { backup => "main" }
</pre>
* Go back and add the line for Inventory Support
<pre>
vi /usr/local/www/puppet-dashboard/config/settings.yml
</pre>
#* And change the following parameters:
<pre>
enable_inventory_service: true
use_file_bucket_diffs: true
</pre>
* With all the updates made, restart so that it takes effect:
<pre>
service nginx restart
</pre>
* For future maintenance, periodic jobs to prune old reports and run DB optimization.
<pre>
mkdir -p /usr/local/etc/periodic/monthly
vi /usr/local/etc/periodic/monthly/clean_dashboard_database.sh
</pre>
#* And add the following:
<pre>
#!/bin/sh
cd /usr/local/share/puppet-dashboard && \
echo "Pruning Old Reports from Puppet Dashboard Database" && \
/usr/bin/su -m puppet-dashboard -c '/usr/local/bin/bundle exec rake RAILS_ENV=production reports:prune upto=3 unit=mon' && \
echo "Optimizing Database" && \
/usr/bin/su -m puppet-dashboard -c '/usr/local/bin/bundle exec rake RAILS_ENV=production db:raw:optimize'
</pre>
#* And make it executable:
<pre>
chmod 755 /usr/local/etc/periodic/monthly/clean_dashboard_database.sh
</pre>
* And create a weekly script:
<pre>
mkdir -p /usr/local/etc/periodic/weekly
vi /usr/local/etc/periodic/weekly/clean_puppet_reports.sh
</pre>
</pre>
#* And add the following:
<pre>
#!/bin/sh
echo "Pruning Puppetmaster Reports greater than 7 days old"
echo -n " Reports Removed:"
find /var/puppet/reports -mtime 7 | xargs rm -v | wc -l
</pre>
#* And make it executable:
<pre>
chmod 755 /usr/local/etc/periodic/weekly/clean_puppet_reports.sh
</pre>
---
h1. Install Icinga2
* Start by installing icinga2 and fcgiwrap:
<pre>
pkg install icinga2
</pre>
*NOTE*: Make sure to uncheck *[ ]MYSQL* during the icinga2 package configuration.
* Start and enable the service to start at boot:
<pre>
echo 'icinga2_enable="YES"' >> /etc/rc.conf
service icinga2 start
</pre>
* Enable the icinga2 ido-pgsql feature:
<pre>
ln -s /usr/local/etc/icinga2/features-available/ido-pgsql.conf /usr/local/etc/icinga2/features-enabled/ido-pgsql.conf
</pre>
* Edit the icinga2 ido-pgsql config:
<pre>
vi /usr/local/etc/icinga2/features-enabled/ido-pgsql.conf
</pre>
#* And modify the following:
<pre>
object IdoPgsqlConnection "ido-pgsql" {
user = "icingauser"
password = "SuperSecretIcingaPassword"
host = "localhost"
database = "icingadb"
}
</pre>
* Load the icinga2 ido-pgsql schema
<pre>
psql -h localhost -U icingauser -W -d icingadb < /usr/local/share/icinga2-ido-pgsql/schema/pgsql.sql
</pre>
* Restart icinga2:
<pre>
service icinga2 restart
</pre>
h2. Install Icingaweb2
* Install icingaweb2:
<pre>
pkg install icingaweb2 php56-pgsql pecl-imagick
</pre>
* Configure the default PHP settings
<pre>
cp /usr/local/etc/php.ini-production /usr/local/etc/php.ini
</pre>
* Set the local timezone:
<pre>
sed -i '' -e 's/;date.timezone\ =/date.timezone\ =\ America\/Los_Angeles/' /usr/local/etc/php.ini
</pre>
* Edit /usr/local/etc/php-fpm.conf:
<pre>
vi /usr/local/etc/php-fpm.conf
</pre>
#* Make the following changes:
<pre>
listen = /var/run/php-fpm.sock
listen.owner = www
listen.group = www
listen.mode = 0660
</pre>
* Start and enable PHP-FPM at boot:
<pre>
echo 'php_fpm_enable="YES"' >> /etc/rc.conf
service php-fpm start
</pre>
* Next, create the *nagios server block*:
<pre>
vi /usr/local/etc/nginx/conf.d/icinga.conf
</pre>
#* And add the following
<pre>
server {
listen 8000 default;
server icinga.example.com;
index index.html index.php;
root /usr/local/www;
# IP and IP ranges which should get access
allow 10.0.0.0/24;
allow 10.1.0.1;
# all else will be denied
deny all;
location ~ ^/icingaweb2/index\.php(.*)$ {
fastcgi_pass unix:/var/run/php-fpm.sock;
fastcgi_index index.php;
include fastcgi_params;
fastcgi_param SCRIPT_FILENAME /usr/local/www/icingaweb2/public/index.php;
fastcgi_param ICINGAWEB_CONFIGDIR /usr/local/etc/icingaweb2;
fastcgi_param REMOTE_USER $remote_user;
}
location ~ ^/icingaweb2(.+)? {
alias /usr/local/www/icingaweb2/public;
index index.php;
try_files $1 $uri $uri/ /icingaweb2/index.php$is_args$args;
}
}
</pre>
* Restart nginx:
<pre>
service nginx restart
</pre>
* Now create a configuration token:
<pre>
cd /usr/local/www/icingaweb2 && ./bin/icingacli setup token create --config=/usr/local/etc/icingaweb2
</pre>
* Enter this token on Icinga Web 2's setup interface at http://localhost:8000/icingaweb2/setup
---
h1. Resources
* http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#configuring-dashboard
* http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#installing-puppet-dashboard
* http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#creating-and-configuring-a-mysql-database
* http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#testing-that-dashboard-is-working
http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#configuring-puppet
* http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#starting-and-managing-delayed-job-workers
* http://docs.puppetlabs.com/dashboard/manual/1.2/bootstrapping.html#running-dashboard-in-a-production-quality-server
* http://docs.puppetlabs.com/puppet/latest/reference/config_important_settings.html
* http://z0mbix.github.io/blog/2012/03/01/use-nginx-and-passenger-to-power-your-puppet-master/
* http://www.watters.ws/mediawiki/index.php/Configure_puppet_master_using_nginx_and_mod_passenger
* http://docs.puppetlabs.com/dashboard/manual/1.2/configuring.html
* https://forums.freebsd.org/viewtopic.php?&t=42071
* http://projects.puppetlabs.com/issues/16686
* http://rlaskey.org/words/897/nagios-nginx-freebsd/
* http://www.unixmen.com/it-appears-as-though-you-do-not-have-permission-to-view-information-for-any-of-the-hosts-you-requested/