Kolla-Ansible One-Click Deployment Guide for OpenStack

Kolla-Ansible is a powerful tool for deploying OpenStack in a containerized manner. It utilizes Docker containers and Ansible automation technology to help you quickly set up a production-grade OpenStack cloud platform.

Below, I will outline the core steps for deployment, key configurations, and some practical suggestions.

📦Understanding the Advantages of Kolla-Ansible

The main reasons for choosing Kolla-Ansible are its outstanding performance in the following areas:

• Containerized Deployment: All OpenStack services run in Docker containers, achieving excellent environment isolation, rapid deployment, and supporting smooth upgrades or rollbacks.
• Automation and Reliability: Utilizing Ansible for automated orchestration reduces manual operation errors, making the deployment process more reliable.
• Community Support: As an official project of the OpenStack community, Kolla-Ansible has good community support and continuous updates and maintenance.

🛠️Preparation Before Deployment

System Requirements

Before starting, please ensure that your target machine (which can be a physical or virtual machine) meets the following basic conditions:

Item	Minimum/Recommended Configuration	Description
Operating System	CentOS 7/8/9, Rocky Linux, Ubuntu 18.04+	Must be a supported Kolla-Ansible Linux distribution. For example, there are reports of deploying the Yoga version on Rocky 8.6 , and there are also cases of deploying the Rocky version on Ubuntu 18.04 .
CPU/Memory	At least 4 cores CPU, 8GB memory	All-in-One (single-node) deployments are also recommended to be configured no lower than this; the more resources, the better.
Disk Space	The system disk must have at least40GB	If you need to enable Cinder block storage service, additional disk space must be prepared for storage volumes.
Network	At least two network interfaces	One for management/API network (usually has an IP address), and another for Neutron external network (usually not configured with an IP).
Virtualization	Support for KVM (recommended) or QEMU	It is recommended to use KVM for better performance on physical machines; testing in virtual machines usually requires setting to QEMU.

System Initialization

1. Disable SELinux and the firewall: To avoid permission and access issues, it is generally recommended to disable SELinux and the firewall in deployment and testing environments.

bash

# Disable SELinux (effective after reboot)

sudo setenforce 0

sudo sed -i ‘s/SELINUX=enforcing/SELINUX=disabled/g’ /etc/selinux/config

# Stop and disable the firewall (Firewalld)

sudo systemctl stop firewalld

sudo systemctl disable firewalld

1. Configure the hostname and hosts file: Set the correct hostname for the node and ensure that the hostname can be correctly resolved to the management network’s IP address in /etc/hosts .

bash

# Set the hostname, for example, set it to ‘controller’

sudo hostnamectl set-hostname controller

# Bind the hostname to the management IP, for example, echo “192.168.41.13 controller” | sudo tee -a /etc/hosts

1. Configure SSH passwordless login: Although single-node deployments (ansible_connection=local) may not require this, multi-node deployments need to configure passwordless SSH login from the deployment node to all target nodes.

🚀Core Deployment Steps

Phase One: Install Dependencies and Kolla-Ansible

1. Install system dependencies:

bash

# For CentOS/Rocky Linux and other RHEL-based systems

sudo dnf install -y python3-devel libffi-devel gcc openssl-devel python3-libselinux libvirt-devel

# Or use yum (for older versions)

# For Ubuntu

sudo apt-get update

sudo apt-get install -y python3-dev libffi-dev gcc libssl-dev

1. Create a Python virtual environment (recommended): Using a virtual environment can avoid Python package version conflicts.

bash

python3 -m venv /path/to/your/venv # For example, /root/venv or /home/username/kolla-venv

source /path/to/your/venv/bin/activate

1. Upgrade pip and install Ansible, Kolla-Ansible:

bash

pip install -U pip

# Install a specific version of Ansible, for example, pip install ‘ansible>=4,<6’:cite[1] or ‘ansible-core>=2.15,<2.16.99’:cite[6]

pip install ‘ansible>=4,<6’

# Install a specific version of Kolla-Ansible, for example, pip install kolla-ansible==14.0.0 (corresponding to Yoga):cite[1]

# Or install directly from the stable branch, for example, pip install git+https://opendev.org/openstack/kolla-ansible@stable/2024.1:cite[6]

pip install kolla-ansible

1. Copy configuration files:

bash

sudo mkdir -p /etc/kolla

sudo chown $USER:$USER /etc/kolla

cp -r /path/to/venv/share/kolla-ansible/etc_examples/kolla/* /etc/kolla/

cp /path/to/venv/share/kolla-ansible/ansible/inventory/* .

Phase Two: Configure Kolla-Ansible

1. Configure Ansible: Create or modify /etc/ansible/ansible.cfg and add the following basic configurations to optimize the experience.

ini

[defaults]

host_key_checking = False

pipelining = True

forks = 100

1. Edit the global configuration file /etc/kolla/globals.yml: This is the most critical configuration step. Below is an example of key configurations for an All-in-One deployment.

yaml

# Basic Configuration

kolla_base_distro: “centos” # or “rocky”, “ubuntu”

kolla_install_type: “binary” #or “source”, usually use binary

openstack_release: “yoga” # Specify the version of OpenStack you want to deploy, such as 2024.1, yoga, wallaby , etc.

# Network Configuration – Modify according to your actual network interface names

network_interface: “ens33” # Management network interface (e.g., the network card used for SSH connections)

neutron_external_interface: “ens34” # External network interface (for virtual machine access to the internet)

kolla_internal_vip_address: “192.168.41.13”# Management network’s VIP, can be set to the management network card’s IP:cite[1]:cite[8]

# Neutron Network Plugin

neutron_plugin_agent: “openvswitch”

# Service Enablement Configuration – Enable as needed

enable_haproxy: “no” # Can disable HAProxy for single-node deployments:cite[1]:cite[3]:cite[8]

enable_cinder: “yes” # Enable block storage service

enable_cinder_backend_lvm: “yes” #Enable LVM backend for Cinder

cinder_volume_group: “cinder-volumes”#Volume group name that needs to be created in advance:cite[4]:cite[6]

# Virtualization Type

nova_compute_virt_type: “qemu” #Use “qemu” when deploying in virtual machines or when physical machines do not support KVM

• Prepare LVM for Cinder (if enabled):

bash

# Assuming the disk device prepared for Cinder is /dev/sdb

sudo pvcreate /dev/sdb

sudo vgcreate cinder-volumes /dev/sdb

1. Configure the Inventory file: For All-in-One deployments, modify the all-in-one inventory file to set all nodes under each group to local connection.

ini

[control]

localhost ansible_connection=local

[network]

localhost ansible_connection=local

[compute]

localhost ansible_connection=local

[storage]

localhost ansible_connection=local

[monitoring]

localhost ansible_connection=local

[deployment]

localhost ansible_connection=local

1. Generate password file:

bash

kolla-genpwd

All service passwords will be generated in /etc/kolla/passwords.yml. You can modify the keystone_admin_password in it before deployment to set the OpenStack admin password.

Phase Three: Execute Deployment

1. Bootstrap Target Servers: Install necessary system packages and configurations.

bash

kolla-ansible -i all-in-one bootstrap-servers

1. Perform Pre-checks: Check if the environment meets deployment requirements.

bash

kolla-ansible -i all-in-one prechecks

1. Pull Docker Images (optional but recommended): Pulling images in advance can speed up deployment.

bash

kolla-ansible -i all-in-one pull

1. Start Deployment: This is the most time-consuming step.

bash

kolla-ansible -i all-in-one deploy

1. Post-deployment Configuration:

bash

# Generate admin authentication file

kolla-ansible post-deploy

# Load environment variables

source /etc/kolla/admin-openrc.sh

# Install OpenStack command-line client

pip install python-openstackclient

🧪Verification and Initial Use

After successful deployment, you can perform the following actions:

• Check Container Status: Run docker ps to see if all OpenStack service containers are running normally.
• Access Horizon Dashboard: Enter the management network’s IP or VIP address in the browser, and log in using the keystone_admin_password corresponding to the password in /etc/kolla/passwords.yml with the admin user.
• Use Command Line: Execute openstack token issue, if it successfully returns token information, it indicates that the CLI is configured correctly.
• Create Sample Network and Instance (optional): You can run source /path/to/venv/share/kolla-ansible/init-runonce to create a demonstration environment.

💡Notes and Troubleshooting

• Version Compatibility: Ensure that the openstack_release configuration is compatible with the version of kolla-ansible you installed.
• Network Interface Configuration: The network interface specified by neutron_external_interface should not be configured with an IP address, otherwise it may affect the virtual machine network.
• Insufficient Resources: When deployment or virtual machine creation fails, check if there is sufficient disk space and memory.
• Check Logs: If a service fails to start, you can use the docker logs <container_name> command to view the corresponding container’s logs to locate the issue.
• Redeploy: If deployment fails, you can first run kolla-ansible -i all-in-one destroy to clean up the environment, and then re-execute the deployment steps.

💎Summary

Kolla-Ansible significantly simplifies the deployment and maintenance of OpenStack through containerization and automation. The key lies in carefully preparing the environment, correctly editing the /etc/kolla/globals.yml configuration file, and following the standard deployment process.

I hope this guide helps you successfully set up your OpenStack cloud platform! If you encounter issues during specific steps, such as a service failing to start, you can refer to the container’s log information.

#OpenStack

#Ansible

#CloudPlatform