Requirements

To install OpenStack is necessary that each node have a Linux operating system Red Hat Enterprise Linux 7 or CentOS 7.

Before any operation, in every node, upgrading packages that already exist with the command :

# yum update -y

Architecture

Before starting with the installation process you need to know from the outset the type of architecture that you want to use for OpenStack.

This project allows you to install OpenStack in HA mode, which means that you can have multiple nodes that have the same functionality within the infrastructure. Obviously you have to have at least one node for each function.

In this guide we will use the following structure :

  • 1 Puppet Master Node
  • 1 Galera Master Node
  • 2 Galera Slave Nodes
  • 3 HA Proxy Nodes

Puppet Master

The Puppet Master is the node that will be tasked to orchestrate the installation of all other nodes of its features automatically.

First you need to install git on the node :

# yum install git

After that you must clone the project directly in the root's home directory:

# cd /root/
# git clone https://github.com/Eleirbag89/prisma.git

Now modify the file /root/prisma/data/common.yaml customizing the following values:

  • DNS
    This is the DNS of Local Network, example: 
    dns : "chainreds-lan"
  • Puppet Master IP
    This is the IP of Puppet Master node: 
    ip_puppet_m : "10.0.0.1"
  • Galera nodes IPs
    This is a list of all IP addresses of Galera nodes: 
    galera_ips : ["10.0.0.2", "10.0.0.3", "10.0.0.4"]
  • Galera nodes Hostnames
    This is a list of all Hostnames of Galera nodes: 
    galera_hosts : ["galera-master", "galera-1", "galera-2"]
  • Galera nodes Wildcard
    This is the IP classes for Galera Nodes: 
    wildcard : "10.0.0.%"
  • Database Password
    This is the database password of Galera Cluster: 
    db_root_password : "password"
  • Disables support for OpenStack
    Disable the HAProxy support for OpenStack: 
    haproxy_monitor_openstack : "false"
  • HA Proxy nodes IPs
    This is a list of all IP addresses of HA Proxy nodes: 
    haproxy_ips : ["10.0.0.5", "10.0.0.6", "10.0.0.7"]
  • HA Proxy nodes Hostnames
    This is a list of all Hostnames of HA Proxy nodes: 
    haproxy_hosts : ["haproxy-1", "haproxy-2", "haproxy-3"]
  • HA Proxy nodes Priorities
    This is a list of all priorities for loadbalance of HA Proxy nodes: 
    haproxy_priority : ["100", "101", "102"]
  • Virtual Public IP
    This is a virtual IP address that will be used at the administrative level between the nodes and HA Proxy to connect to the Database: 
    ip_hap_v : "192.168.7.100"
  • Virtual Internal IP
    This is a virtual IP address that will be used to communicate between the nodes of OpenStack: 
    ip_hap_v_private : "10.0.0.100"
  • Network Interface for the Virtual Public IP
    The name of interface can comunicate with the Virtual Public IP: 
    vip_interface : "eth0"
  • Network Interface for the Virtual Internal IP
    The name of interface can comunicate with the Virtual Internal IP: 
    vip_interface : "eth1"

At this point we have completed the general configuration of our structure . Before proceeding make sure that the hostname of the Puppet Master is puppet :

# sysctl kernel.hostname=puppet

After you've completed the previous stages, run this script in order to install the dependencies, copy files to their working directories and install a production-ready web server (Apache Passenger) that will run as Puppet Master:

# sh /root/prisma/scripts/centos_install.sh

At the end of the installation process we proceed with the installation of various Puppet Agent

Puppet Agents

Once configured the Puppet Master, this will be used to auto configure all nodes related to our structure.

It's necessary that each node is present Puppet Agent Configuration File and run as root :

# sudo -i
# wget https://raw.githubusercontent.com/Eleirbag89/prisma/gh-pages/scripts/agent_init.sh

Galera Nodes

A Galera node is a database MariaDB Galera. With them you can create a cluster in order to always have a high availability of data.

In a Galera Cluster we can always identify a Master node and many nodes Slave

First we need to install the Master Galera using the uppet Agent Configuration File:

# sudo -i
# sh agent_init.sh

During the installation will be required of the information that shall match those included previously in common.yaml:

-----Inizializzazione nodo di Puppet Agent-----
Inserisci l'indirizzo IP del Puppet Master:10.0.0.1
Inserisci l'Hostname del Puppet Master [default puppet]:
Inserisci l'FQDN del Puppet Master:chainreds-lan
Inserisci l'indirizzo IP di questo Nodo:10.0.0.2
Inserisci l'Hostname di questo Nodo:galera-master
Inserisci l'FQDN di questo Nodo [default chainreds-lan] :
Inserisci il nome dell'interfaccia di rete [default eth1]:

At the end of the installation it will be able to install other Galera nodes in exactly the same way.

HA Proxy Nodes

HAProxy is a free, very fast and reliable solution offering high availability, load balancing, and proxying for TCP and HTTP-based applications. It also allows you to assign Virtual IPs for cluster nodes.

Then we proceed to the installation on every HA Proxy node:

# sudo -i
# sh agent_init.sh

During the installation will be required of the information that shall match those included previously in common.yaml:

-----Inizializzazione nodo di Puppet Agent-----
Inserisci l'indirizzo IP del Puppet Master:10.0.0.1
Inserisci l'Hostname del Puppet Master [default puppet]:
Inserisci l'FQDN del Puppet Master:chainreds-lan
Inserisci l'indirizzo IP di questo Nodo:10.0.0.5
Inserisci l'Hostname di questo Nodo:haproxy-1
Inserisci l'FQDN di questo Nodo [default chainreds-lan] :
Inserisci il nome dell'interfaccia di rete [default eth1]:

Verify Operations

After installation of each HAProxy node we can verify proper operation of the Galera and HA Proxy nodes.

To do this simply open a web browser and go to http://192.168.7.100:8080 and see the Galera Cluster table. If all nodes are green means that all is well:

Add more Nodes

To add a new node, you must first edit the file /etc/puppet/data/common.yaml and /etc/puppet/data/openstack.yaml in Puppet Master.

If the node that you want to add requires use HAProxy you must also upgrade those nodes .

Add Galera Node

After changing adequately file /etc/puppet/data/common.yaml in Puppet Master adding IP and hostname of the new node, you can install the Service with the classic mode:

# sudo -i
# wget https://raw.githubusercontent.com/Eleirbag89/prisma/prisma2/scripts/agent_init.sh
# sh agent_init.sh

After you will have to update the information in the other nodes related to the new one. To do this, in each HAProxy and Galera node execute the following command:

# puppet agent

Add HAProxy Node

After changing adequately file /etc/puppet/data/common.yaml in Puppet Master adding IP, hostname and priority of the new node, you can install the Service with the classic mode:

# sudo -i
# wget https://raw.githubusercontent.com/Eleirbag89/prisma/prisma2/scripts/agent_init.sh
# sh agent_init.sh

After you will have to update the information in the other nodes related to the new one. To do this, in each HAProxy node execute the following command:

# puppet agent