How to Seamlessly Scale Your OpenStack Cloud: Adding Compute Nodes with Kolla-Ansible
As your cloud environment grows, the need for more computational power becomes inevitable. Scaling your OpenStack infrastructure is a critical task that, if done incorrectly, can lead to downtime and configuration headaches. Fortunately, for those who deploy and manage their cloud with Kolla-Ansible, expanding your resource pool by adding new compute nodes is a remarkably streamlined and efficient process.
This guide provides a clear, step-by-step walkthrough for adding a new compute node to an existing Kolla-Ansible-based OpenStack deployment. By leveraging the power of Ansible’s automation, you can ensure a consistent and repeatable scaling strategy for your private cloud.
Prerequisites for Expansion
Before you begin, ensure you have the following in place:
- An Existing OpenStack Deployment: You must have a functioning OpenStack cloud that was initially deployed using Kolla-Ansible.
- A New Server: A new physical or virtual server is ready to be provisioned as the compute node. This server should have a fresh, supported OS installation (like CentOS or Ubuntu) and network connectivity.
- SSH Access: You need passwordless SSH access from your Kolla-Ansible control host (where you run the Ansible commands) to the new server, using an account with
sudoprivileges.
Step-by-Step: Adding the New Compute Node
The process hinges on updating your inventory file to make Ansible aware of the new hardware and then running the appropriate playbooks to configure it.
Step 1: Prepare the New Compute Host
First, ensure your new server is properly prepared. This includes basic system setup:
- Assigning a static IP address.
- Configuring hostname and DNS settings.
- Ensuring the network interfaces are correctly configured to connect to your management, tenant, and storage networks as required by your OpenStack architecture.
- Crucially, verify that you can SSH from the Kolla-Ansible control node to the new server without a password, as this is how Ansible will connect and perform its tasks.
Step 2: Update the Kolla-Ansible Inventory File
The inventory file is the heart of your Kolla-Ansible configuration. It defines the role of every server in your cloud. You must add your new server to this file so Kolla-Ansible knows what to do with it.
- Locate your inventory file. This is typically found at
/etc/kolla/hostsor within your Kolla-Ansible checkout directory (e.g.,ansible/inventory/hosts). - Open the file for editing.
- Find the
[compute]group. - Add the hostname or IP address of your new server to this group.
For example, if your new server is named new-compute-01 with an IP of 192.168.1.15, your [compute] group might look like this:
[compute]
existing-compute-01
existing-compute-02
new-compute-01 ansible_host=192.168.1.15
Adding the ansible_host variable is a best practice if your DNS is not yet configured or you prefer to use IP addresses for Ansible connections.
Step 3: Run the Kolla-Ansible Playbooks
With the inventory updated, it’s time to let Ansible do the heavy lifting. We will run two key commands.
First, you need to bootstrap the new server. This playbook installs essential dependencies like Docker and other required packages. To avoid re-configuring your existing nodes, use the --limit flag to target only the new server.
kolla-ansible -i /path/to/your/inventory bootstrap-servers --limit new-compute-01
Once the bootstrap process is complete, you can deploy the OpenStack compute services to the new node. Again, using a tag or limit is highly recommended to avoid unnecessary work on other nodes. The nova tag is perfect for this, as it includes all services related to the Nova compute project.
kolla-ansible -i /path/to/your/inventory deploy --tags nova
This command will install, configure, and start all the necessary Docker containers for the nova-compute service and its dependencies on new-compute-01.
Step 4: Verify the New Node is Active
After the playbooks complete successfully, the final step is to verify that the OpenStack control plane recognizes the new compute node and that it is ready to host virtual machines.
Log in to a host with the OpenStack client tools installed and run the following command:
openstack compute service list
You should see your new host (new-compute-01 in our example) in the output list. Confirm that the nova-compute service on the new host has a status of enabled and a state of up.
+----+----------------+-------------------+----------+---------+-------+----------------------------+
| ID | Binary | Host | Zone | Status | State | Updated At |
+----+----------------+-------------------+----------+---------+-------+----------------------------+
| 7 | nova-conductor | controller-01 | internal | enabled | up | 2023-10-27T18:30:05.000000 |
| 8 | nova-scheduler | controller-01 | internal | enabled | up | 2023-10-27T18:30:08.000000 |
| 9 | nova-compute | existing-compute-01| nova | enabled | up | 2023-10-27T18:30:07.000000 |
| 10 | nova-compute | existing-compute-02| nova | enabled | up | 2023-10-27T18:30:09.000000 |
| 11 | nova-compute | new-compute-01 | nova | enabled | up | 2023-10-27T18:30:06.000000 |
+----+----------------+-------------------+----------+---------+-------+----------------------------+
Once you see this, your new compute node is fully integrated into the cluster and the OpenStack scheduler can begin placing new instances on it.
Security and Best Practices
- Run Pre-Checks: Before deploying, it’s wise to run
kolla-ansible prechecksto catch potential issues with your configuration or new host. - Use Ansible Vault: For production environments, always store sensitive data like passwords in Ansible Vault rather than in plain text files. This protects your credentials and enhances the security of your cloud.
- Understand Idempotency: Kolla-Ansible playbooks are idempotent, meaning you can safely re-run them. If a deployment fails midway, you can often fix the underlying issue and simply run the command again without negative side effects.
By following this structured approach, you can confidently and safely scale your OpenStack cloud’s compute capacity, ensuring your infrastructure keeps pace with your organization’s demands.
Source: https://kifarunix.com/add-compute-nodes-into-openstack-using-kolla-ansible/
