🚀 OpenStack Single-Node Deployment using Kolla Ansible

Based on Official Guide: Kolla Ansible 2025.1 Development Quickstart

✅ All-in-One (AIO) Setup | ✅ VirtualBox VM | ✅ Fixed Common Issues

✅ Perfect for Evaluation, Learning Development & Production
🔧 Deploy OpenStack All-in-One with Kolla Ansible — Fast, Reliable & Professional

💡 About Kolla Ansible

Kolla Ansible is an official OpenStack project that uses Ansible to deploy containerized OpenStack services via Docker. It’s ideal for:

Rapid deployment of OpenStack for testing or evaluation.
Consistent, repeatable infrastructure as code.
Production-ready architecture (scalable from single-node to multi-node).

✅ Why Use Kolla Ansible?

Fully automated setup

Built-in high availability (VIP + keepalived)

Easy upgrades and maintenance

Supports Ubuntu, Rocky Linux, CentOS

This guide uses the development-style installation method (cloning from Git) as per the official Kolla Ansible quickstart.

⚙️ Prerequisites

Ensure your VM meets these requirements:

Requirement	Details
OS	Ubuntu 24.04 LTS or Rocky Linux 9
RAM	8 GB or more
Disk	40+ GB free space
Network Interfaces	3 NICs (NAT, Host-Only, Bridged)
User Privileges	Root access or sudo rights

🖥️ Our Environment:

✅ 2+ vCPUs

✅ 8+ GB RAM

✅ 40+ GB Disk

✅ OS: Ubuntu 24.04.X LTS

✅ 3 Network Interfaces (as described)

Interface	Type	IP Address	Purpose
`enp0s3`	NAT	`10.0.2.15`	Internet access
`enp0s8`	Host-Only	`192.168.10.10`	Management network
`enp0s9`	Bridged	`192.168.102.221`	External network for instances

We'll use:

enp0s8 → network_interface
enp0s9 → neutron_external_interface
VIP → 192.168.10.250

Run all commands as root or with sudo.

➕ Step-by-Step Installation

💡 Run all commands as root or use sudo.

1. Update System & Install Dependencies

Update package list

sudo apt update && sudo apt upgrade -y

Install system build tools and Python dependencies

apt install -y git python3-dev libffi-dev gcc libssl-dev libdbus-glib-1-dev python3-venv

🐧 For Rocky/CentOS, use:

dnf install -y git python3-devel libffi-devel gcc openssl-devel python3-libselinux

2. Create Virtual Environment

python3 -m venv /opt/venv-kolla
source /opt/venv-kolla/bin/activate
pip install --upgrade pip
pip install docker
pip install dbus-python

🔁 Keep this environment activated for all steps!

3. Clone & Install Kolla-Ansible from Git

git clone --branch stable/2025.1 https://opendev.org/openstack/kolla-ansible

Install Python dependencies

pip install ./kolla-ansible

4. Create Config Directory

mkdir -p /etc/kolla
chown $USER:$USER /etc/kolla

5. Copy Configuration Files

cp -r kolla-ansible/etc/kolla/* /etc/kolla/
cp kolla-ansible/ansible/inventory/* .

6. Install Ansible Galaxy Dependencies

kolla-ansible install-deps

7. Generate Passwords

kolla-ansible/tools/generate_passwords.py

🔐 All passwords are now securely generated in /etc/kolla/passwords.yml

8. Configure `globals.yml`

Edit the main config file:

nano /etc/kolla/globals.yml

Update with your settings:

# Base OS for containers
kolla_base_distro: "ubuntu"

# OpenStack release
kolla_install_type: "source"

# Management interface (host-only)
network_interface: "enp0s8"

# External interface for Neutron (bridged, no IP!)
neutron_external_interface: "enp0s9"

# Virtual IP for internal traffic (must be unused)
kolla_internal_vip_address: "192.168.10.250"

# Enable key services
enable_haproxy: "yes"
enable_keystone: "yes"
enable_glance: "yes"
enable_nova: "yes"
enable_neutron: "yes"
enable_placement: "yes"
enable_horizon: "yes"
enable_cinder: "yes"
enable_cinder_backend_lvm: "yes"
enable_heat: "yes"

Save and exit (Ctrl+O, Enter, Ctrl+X).

9. Bootstrap the Server

Installs Docker and other required tools.

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

10. Run Pre-Deployment Checks

kolla-ansible prechecks -i all-in-one

❌ If it fails due to missing Python modules, see Troubleshooting below.

11. Deploy OpenStack

Deploy all services in containers:

kolla-ansible deploy -i all-in-one

🕐 This may take 10–20 minutes.

✅ Wait for: PLAY RECAP with 0 failed tasks.

12. Generate Admin Credentials

kolla-ansible post-deploy -i ~/all-in-one

✅ Output: /etc/kolla/clouds.yaml and /etc/kolla/admin-openrc.sh

13. Install OpenStack CLI

pip install python-openstackclient -c https://releases.openstack.org/constraints/upper/2025.1

14. Source Admin Environment

source /etc/kolla/admin-openrc.sh

Test connection:

openstack token issue

✅ You should see a valid authentication token.

15. (Optional) Run Demo Script

Creates sample network, image, and instance.

kolla-ansible/tools/init-runonce

⚠️ Only for testing! Do not run in production.

🔍 Verify & Use OpenStack

Access Horizon Dashboard

Open in browser:

http://192.168.10.10
or
http://192.168.102.221

Username: admin
Password: See /etc/kolla/passwords.yml → keystone_admin_password

Get password:

grep keystone_admin_password /etc/kolla/passwords.yml

🛠 Manage & Maintain Your Cloud

Useful Commands

Task	Command
Redeploy (e.g., after config change)	`kolla-ansible deploy -i ~/all-in-one`
Restart specific service (Nova)	`kolla-ansible deploy -i ~/all-in-one --tags nova`
Stop all containers	`kolla-ansible stop -i ~/all-in-one`
Pull latest images	`kolla-ansible pull -i ~/all-in-one`
View logs	`docker logs <container_name>`
List running containers	`docker ps`

Upgrade OpenStack (Future)

When upgrading to a new version:

# Pull new images
kolla-ansible pull -i ~/all-in-one

# Run upgrade playbook
kolla-ansible upgrade -i ~/all-in-one

⚠️ Always backup first!

🧹 Cleanup / Uninstall

To completely remove OpenStack:

kolla-ansible destroy -i ~/all-in-one --yes-i-really-really-mean-it

⚠️ Warning: This deletes all data, including volumes and databases.

Then remove config and virtual env:

rm -rf /etc/kolla
rm -rf /opt/venv-kolla
rm -rf ~/kolla-ansible

🔧 Troubleshooting & Common Issues

Here are real problems you faced and their perfect fixes, plus other likely issues.

❌ Issue 1: `ModuleNotFoundError: No module named 'docker'`

Error during prechecks:

failed: [localhost] ... ModuleNotFoundError: No module named 'docker'

✅ Solution:

Install docker in the Ansible runtime environment:

pip install docker

/opt/ansible-runtime/bin/pip install docker

❌ Issue 2: `ModuleNotFoundError: No module named 'dbus'`

Error:

import dbus; ModuleNotFoundError: No module named 'dbus'

✅ Solution:

Install dbus-python in the Ansible runtime:

pip install dbus-python

/opt/ansible-runtime/bin/pip install dbus-python

🔗 Requires system package: libdbus-glib-1-dev (already installed above).

❌ Issue 3: `Inventory path does not exist` during `post-deploy`

Error:

Kolla inventory /etc/kolla/ansible/inventory/all-in-one is invalid: Path does not exist

✅ Solution:

Create the correct directory and copy the file:

mkdir -p /etc/kolla/ansible/inventory
cp all-in-one /etc/kolla/ansible/inventory/all-in-one

Or always specify -i:

kolla-ansible post-deploy -i all-in-one

❌ Issue 4: Neutron external network not working

Symptoms:

Instances can't reach internet
Floating IPs not accessible

✅ Fix:

Ensure neutron_external_interface has no IP address:

ip addr flush dev enp0s9

And disable any network manager control over it.

❌ Issue 5: Time synchronization failure

Error in prechecks about NTP.

✅ Fix:

Enable time sync:

timedatectl set-ntp true
timedatectl status

❌ Issue 6: Not enough memory or disk space

Symptom: Docker fails to start or containers crash.

✅ Fix:

Allocate at least 8GB RAM and 40GB disk to VM.
Monitor usage: df -h, free -h

❌ Issue 7: Fixing "Volume group cinder-volumes not found" Error

1. Error Overview

Error Message:

fatal: [localhost]: FAILED! => {"changed": false, "cmd": ["vgs", "cinder-volumes"], ... 
"stderr": "Volume group \"cinder-volumes\" not found"}

When it happens: During kolla-ansible prechecks or deploy.

2. Root Cause

Default Behavior: OpenStack Cinder uses LVM storage by default.
Requirement: It looks for a specific Volume Group named cinder-volumes.
Your Setup: You have only one disk (used for OS). No extra disk exists for Cinder.
Result: Precheck fails because the storage group is missing.

3. Solution: Create a Loopback Device

We will create a large file on your existing disk and treat it as a virtual hard disk for Cinder.

Step 1: Create a Virtual Disk File

Create a 20GB file named cinder-volumes.img in the root directory.

dd if=/dev/zero of=/cinder-volumes.img bs=1M count=20480

Step 2: Setup Loop Device

Connect the file to a loop device (e.g., /dev/loop0).

losetup /dev/loop0 /cinder-volumes.img

Step 3: Create Physical Volume (PV)

Initialize the loop device for LVM.

pvcreate /dev/loop0

Step 4: Create Volume Group (VG)

Create the required group named cinder-volumes.

vgcreate cinder-volumes /dev/loop0

4. Verification

Check if VG exists:

vgs

You should see cinder-volumes in the list.

Re-run Prechecks:

kolla-ansible prechecks -i all-in-one

The Cinder error should be gone.

5. Important Notes

A. Persistence (After Reboot)

The loop device will disconnect after reboot. To fix this automatically:

Open /etc/rc.local (create if not exists).

Add these lines before exit 0:

losetup /dev/loop0 /cinder-volumes.img
vgchange -ay

Make sure /etc/rc.local is executable.

B. Storage Capacity

Total Size: 20 GB.
Usable Size: ~18 GB (some space used for LVM metadata).
Usage: You can create multiple volumes (e.g., 10 volumes of 1GB each). Once full, you cannot create new volumes.

C. Performance

Lab/Testing: Perfectly fine.
Production: Not recommended. It is slower than a physical disk.

D. Future Migration to Ceph

If you add Ceph storage later:

Update globals.yml to enable cinder_backend_ceph.
Disable cinder_backend_lvm.
Re-run kolla-ansible deploy.
You do not need to delete the loopback file, but Cinder will stop using it for new volumes.

❌ Issue 7: Nova Compute Failed to Register (Libvirt SASL Authentication Error)

🔴 Error Title

"The Nova compute service failed to register itself"
(Caused by: Libvirt SASL Authentication Failed)

📋 Error Description

When running kolla-ansible deploy, the playbook fails at the nova-cell role with this message:

fatal: [localhost]: FAILED! => {"msg": "The Nova compute service failed to register itself on the following hosts: controller"}

Related log errors from nova-compute.log:

nova.exception.HypervisorUnavailable: Connection to the hypervisor is broken on host
libvirt.libvirtError: authentication failed: authentication failed

What this means:
The nova-compute container cannot connect to the nova-libvirt container because the SASL (Simple Authentication and Security Layer) password verification is failing.

🎯 Root Causes

Cause	Explanation
1. Corrupted passwd.db file	The `/etc/libvirt/passwd.db` file (which stores SASL passwords) got corrupted or has wrong format. This is a binary file, and sometimes it gets broken during container startup or config copy.
2. Password mismatch	The password in `nova-compute`'s `auth.conf` does not match the password in `nova-libvirt`'s `passwd.db`.
3. Container startup timing	`nova-compute` tries to connect before `nova-libvirt` is fully ready. This is a race condition.
4. Incomplete config propagation	After updating `passwords.yml`, the new password was not properly copied to the containers.

✅ Solutions

🔹 Quick Fix (For Lab/Testing Only)

Disable SASL authentication temporarily to complete deployment:

Edit /etc/kolla/globals.yml:

libvirt_enable_sasl: false
nova_libvirt_enable_sasl: false

Re-deploy only nova services:

kolla-ansible deploy -i <inventory> --limit controller --tags nova

⚠️ Warning: Only use this for development or testing. Do not use in production.

🔹 Permanent Fix (For Production)

Fix the SASL authentication properly:

Step 1: Verify passwords.yml

grep libvirt_sasl /etc/kolla/passwords.yml

Make sure you see:

libvirt_sasl_username: nova
libvirt_sasl_password: <your_password>

Step 2: Reset passwd.db in nova_libvirt container

#### Enter the container
docker exec -it nova_libvirt bash

#### Remove corrupted file
rm -f /etc/libvirt/passwd.db

#### Create new clean file
touch /etc/libvirt/passwd.db
chmod 640 /etc/libvirt/passwd.db
chown root:nova /etc/libvirt/passwd.db

#### Add correct password entry (use your actual password)
SASL_PASS="your_password_here"
saslpasswd2 -p -c -f /etc/libvirt/passwd.db -u openstack nova <<< "$SASL_PASS"

#### Verify (use strings, not cat)
strings /etc/libvirt/passwd.db
> Expected output: nova@openstack:your_password_here

#### Exit container
exit

Step 3: Verify auth.conf in nova_compute container

docker exec -it nova_compute bash
cat /var/lib/nova/.config/libvirt/auth.conf

Make sure the password value matches the one in passwd.db.

Step 4: Restart containers in correct order

docker restart nova_libvirt
sleep 10
docker restart nova_compute

Step 5: Verify the fix

# Check nova-compute logs
docker logs -tf nova_compute | grep -E "Connected|ERROR"

# Check OpenStack service status
source /etc/kolla/admin-openrc.sh
openstack compute service list --service nova-compute

✅ Success: Status should show up and :-)

🔍 Verification Checklist

Run these commands to confirm everything is working:

# 1. Check container status
docker ps | grep -E "nova_compute|nova_libvirt"

# 2. Check passwd.db content (use strings)
docker exec nova_libvirt strings /etc/libvirt/passwd.db

# 3. Check auth.conf content
docker exec nova_compute cat /var/lib/nova/.config/libvirt/auth.conf

# 4. Test libvirt connection
docker exec nova_compute virsh -c qemu+tcp://controller/system list

# 5. Check OpenStack service
source /etc/kolla/admin-openrc.sh
openstack compute service list --service nova-compute

🛡️ Prevention Tips

Tip	Why It Helps
✅ Always run full `kolla-ansible deploy` after changing `passwords.yml`	Ensures all configs are properly copied to containers
✅ Use `strings` command to check `passwd.db`	`cat` shows garbage for binary files; `strings` shows readable content
✅ Restart `nova_libvirt` before `nova_compute`	Avoids race conditions during startup
✅ Keep `globals.yml` and `passwords.yml` in sync	Prevents password mismatch errors
✅ Test SASL disabled first in lab, then enable for production	Easier debugging during development

📊 Environment Recommendation

Environment	SASL Setting	Reason
🧪 Lab / Learning	`false`	Faster setup, easier debugging
🏢 Production	`true`	Required for security
🔄 Hybrid (Test → Prod)	Start `false`, then `true`	Test features first, then secure

🆘 Still Having Issues?

If the problem persists, collect this information for further help:

# 1. Nova compute logs (last 50 lines)
docker logs --tail 50 nova_compute

# 2. SASL file content
docker exec nova_libvirt strings /etc/libvirt/passwd.db

# 3. Auth config
docker exec nova_compute cat /var/lib/nova/.config/libvirt/auth.conf

# 4. Libvirt network status
docker exec nova_libvirt netstat -tlnp | grep 16509

Share these outputs with your support team or community for faster troubleshooting.

💡 Final Note: If you disabled SASL for testing and your deployment is now working, your OpenStack cloud is functional. You can enable SASL later when you are ready for production hardening.

📜 License & Credits

Guide Version: 2025.1
Author: Sumon Paul
Source: Based on OpenStack Kolla Ansible Docs
License: MIT (Feel free to reuse and modify)

✅ Congratulations!
You’ve successfully deployed a fully functional single-node OpenStack cloud using Kolla Ansible.

Use this environment to learn Nova, Neutron, Cinder, Horizon, and more!

📬 Need Help?
Have questions or improvements? Open an issue or contact me!

✨ Happy Cloud Building!

FilesExpand file tree

openstack-aio-v2.md

Latest commit

History

openstack-aio-v2.md

File metadata and controls

🚀 OpenStack Single-Node Deployment using Kolla Ansible

Based on Official Guide: Kolla Ansible 2025.1 Development Quickstart

📌 Table of Contents

💡 About Kolla Ansible

⚙️ Prerequisites

🖥️ Our Environment:

➕ Step-by-Step Installation

1. Update System & Install Dependencies

Update package list

Install system build tools and Python dependencies

2. Create Virtual Environment

3. Clone & Install Kolla-Ansible from Git

Install Python dependencies

4. Create Config Directory

5. Copy Configuration Files

6. Install Ansible Galaxy Dependencies

7. Generate Passwords

8. Configure globals.yml

9. Bootstrap the Server

10. Run Pre-Deployment Checks

11. Deploy OpenStack

12. Generate Admin Credentials

13. Install OpenStack CLI

14. Source Admin Environment

15. (Optional) Run Demo Script

🔍 Verify & Use OpenStack

Access Horizon Dashboard

🛠 Manage & Maintain Your Cloud

Useful Commands

Upgrade OpenStack (Future)

🧹 Cleanup / Uninstall

🔧 Troubleshooting & Common Issues

❌ Issue 1: ModuleNotFoundError: No module named 'docker'

❌ Issue 2: ModuleNotFoundError: No module named 'dbus'

❌ Issue 3: Inventory path does not exist during post-deploy

❌ Issue 4: Neutron external network not working

❌ Issue 5: Time synchronization failure

❌ Issue 6: Not enough memory or disk space

❌ Issue 7: Fixing "Volume group cinder-volumes not found" Error

1. Error Overview

2. Root Cause

3. Solution: Create a Loopback Device

Step 1: Create a Virtual Disk File

Step 2: Setup Loop Device

Step 3: Create Physical Volume (PV)

Step 4: Create Volume Group (VG)

4. Verification

5. Important Notes

A. Persistence (After Reboot)

B. Storage Capacity

C. Performance

D. Future Migration to Ceph

❌ Issue 7: Nova Compute Failed to Register (Libvirt SASL Authentication Error)

🔴 Error Title

📋 Error Description

🎯 Root Causes

✅ Solutions

🔹 Quick Fix (For Lab/Testing Only)

🔹 Permanent Fix (For Production)

Step 1: Verify passwords.yml

Step 2: Reset passwd.db in nova_libvirt container

Step 3: Verify auth.conf in nova_compute container

Step 4: Restart containers in correct order

Step 5: Verify the fix

🔍 Verification Checklist

🛡️ Prevention Tips

📊 Environment Recommendation

🆘 Still Having Issues?

📜 License & Credits

8. Configure `globals.yml`

❌ Issue 1: `ModuleNotFoundError: No module named 'docker'`

❌ Issue 2: `ModuleNotFoundError: No module named 'dbus'`

❌ Issue 3: `Inventory path does not exist` during `post-deploy`