kolla-ansible-train

Deploy a openstack demo env of train version via kolla-ansbile step-by-step.

My Lab Environment

To deploy OpenStack using Kolla Ansible, we will be using VMs created using VMware Workstation. So, I have already installed VMware Workstation on my Windows Desktop on which I have created three VMs using the following specs:

Name: control01
OS: Ubuntu 18.04.6 LTS
RAM: 8GB
Disk: 40 GB
CPU: Quad Core
ens33(NAT): 10.0.0.11
ens34(Host-only): Make sure no IP Address is assigned to this interface

Name: compute01
OS: Ubuntu 18.04.6 LTS
RAM: 4GB
Disk: 40 GB
CPU: Quad Core
ens33(NAT): 10.0.0.31
ens34(Host-only): Make sure no IP Address is assigned to this interface

Name: storage01
OS: Ubuntu 18.04.6 LTS
RAM: 4GB
Disk: 40 GB
Disk2: 40 GB (Addtional disk /dev/sdb for creating cinder-volumes volume group)
CPU: Quad Core
ens33(NAT): 10.0.0.41

Install and Setup Kolla Ansible

Generates a pip conf on all of the nodes, creates vloume group on storage, and then install dependency packages using virtual environment on deployment before we go ahead and install Kolla Ansible:

ansible-playbook -i hosts site.yml

Edit multinode

# These initial groups are the only groups required to be modified. The
# additional groups are for more control of the environment.
[control]
# These hostname must be resolvable from your deployment host
control01   ansible_host=10.0.0.11  ansible_python_interpreter=/usr/bin/python3

# The above can also be specified as follows:
#control[01:03]     ansible_user=kolla

# The network nodes are where your l3-agent and loadbalancers will run
# This can be the same as a host in the control group
[network]
control01

[compute]
compute01   ansible_host=10.0.0.31  ansible_python_interpreter=/usr/bin/python3

[monitoring]
control01

# When compute nodes and control nodes use different interfaces,
# you need to comment out "api_interface" and other interfaces from the globals.yml
# and specify like below:
#compute01 neutron_external_interface=eth0 api_interface=em1 storage_interface=em1 tunnel_interface=em1

[storage]
storage01   ansible_host=10.0.0.41  ansible_python_interpreter=/usr/bin/python3

[deployment]
localhost       ansible_connection=local      ansible_python_interpreter=/usr/bin/python3

[baremetal:children]
control
network
compute
storage
monitoring

Configure globals.yml

---
kolla_base_distro: "ubuntu"
kolla_install_type: "binary"
openstack_release: "train"
kolla_internal_vip_address: "10.0.0.20"
docker_registry: "registry.cn-guangzhou.aliyuncs.com"
docker_registry_username: "aliyun3882791286"
docker_namespace: "lunwei"
network_interface: "ens33"
neutron_external_interface: "ens34"
enable_cinder: "yes"
enable_cinder_backend_lvm: "yes"
nova_compute_virt_type: "qemu"

Check Inventory

source venv/bin/activate
ansible -i multinode all -m ping

Generate Password

In the next step generate the password for the services. These generated passwords will be stored in /etc/kolla/passwords.yml file

kolla-genpwd

Modify Password

For convenience, change the keystone_admin_password which is used to login in dashboard. By the way, adding docker_registry_password for docker login:

sed -i 's#keystone_admin_password:.*#keystone_admin_password: kolla#g' /etc/kolla/passwords.yml
sed -i 's#docker_registry_password:.*#docker_registry_password: my_password#g' /etc/kolla/passwords.yml

Perform Deployment

Now in the next step we are going to install bootstrap servers.

kolla-ansible -i multinode bootstrap-servers

Always do deployment checks.

kolla-ansible -i multinode prechecks

In the last step we will perform deployment.

kolla-ansible -i multinode deploy

Now install curl and the Openstack client tools

sudo apt install -y curl python-openstackclient

Upload images to private hub (optional)

You can pull images before depoly:

kolla-ansible pull

Add local private hub to docker config:

sudo tee /etc/docker/daemon.json <<-'EOF'
{
    "insecure-registries": [
        "10.0.0.10:4000"
    ],
    "log-opts": {
        "max-file": "5",
        "max-size": "50m"
    }
}
EOF
sudo systemctl daemon-reload
sudo systemctl restart docker

To upload these images to local private hub:

for i in $(sudo docker images | grep train | awk '{print $1":"$2}')
do
  bash auto_image_push.sh $i
done

update /etc/kolla/globals.yml as below if you use a local private hub to deploy:

---
kolla_base_distro: "ubuntu"
kolla_install_type: "binary"
openstack_release: "train"
kolla_internal_vip_address: "10.0.0.20"
docker_registry: "10.0.0.10:4000"
docker_namespace: "kolla"
network_interface: "ens33"
neutron_external_interface: "ens34"
enable_cinder: "yes"
enable_cinder_backend_lvm: "yes"
nova_compute_virt_type: "qemu"

Post Deployment

Now create admin-openrc.sh file that basically contains admin credentials.

kolla-ansible post-deploy /etc/kolla/admin-openrc.sh

Now in the next step we will run the deployment.

cp /etc/kolla/admin-openrc.sh .
source admin-openrc.sh

The initialization script will create virtual machine resources such as cirros image, network, subnet, routing, security group, specification, quota, etc.

cd venv/share/kolla-ansible
./init-runonce

Navigate http://10.0.0.20 to login into the dashboard of OpenStack.

username: admin
password: kolla