--- title: "Lifecycle of an OPCP Node" description: "Discover the lifecycle of an OPCP node and its different statuses" url: https://docs.ovhcloud.com/en/guides/hosted-private-cloud/opcp/node-lifecycle lang: en lastUpdated: 2025-11-18 --- # Lifecycle of an OPCP Node ## Objective A node in OpenStack represents the configuration of a physical server in the OPCP rack. It must be distinguished from instances, which represent the operating system running on a node. **This guide details the different statuses of a node in an OPCP rack and how to modify them.** ## Requirements - Have an active [OPCP](https://www.ovhcloud.com/en-gb/hosted-private-cloud/onprem-cloud-platform/) service. - Have a user account with admin rights to log in to Horizon on the OPCP offering. - (Optional) Have access to the OpenStack APIs for your project. - (Optional) Have installed the Ironic client. ## Instructions Log in to the Horizon interface of your OPCP on the admin project. ![dashboard](/images/hosted-private-cloud/opcp/node-lifecycle/01-log-to-horizon.png) If you want to follow the OpenStack API section, you will need to install the Ironic packages on your environment: ```bash pip install python-ironicclient ``` ### Check the status of a node You can check the status of a node directly from Horizon via the `Admin` > `System` > `Ironic Bare Metal Provisioning` tab: ![server-status](/images/hosted-private-cloud/opcp/node-lifecycle/02-server-status.png) You will find the list of your nodes and their various statuses. From the OpenStack APIs, you can retrieve the same list using the following command: ```bash baremetal node list ``` You can also check the status of a specific node: ```bash baremetal node show $BAREMETAL_NODE_ID ``` ### Possible statuses | Status | Description | | -------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Enroll | First state of the node when it has been automatically discovered by OPCP. The server has not yet been validated and must be manually transitioned to `Manageable`. | | Manageable | The node has been verified and is managed by Ironic, but it is not yet installable. The node must be moved to the `Available` state before it can be deployed. | | Available | The node is available and can be installed. | | Active | The node is installed and has an active instance on it. | | Verifying | Transitional state when a node moves from `Enroll` to `Manageable`. Ironic verifies it can manage the node using the drivers and hardware properties configured during control-plane discovery. | | Cleaning / Clean-wait | Transitional state when an instance is deleted or when leaving the `Manageable` state before becoming `Available` again. Disks are wiped during this step. | | Deploying / Wait call-back | Transitional state when the node is being deployed. | You can find detailed explanations for the different statuses in the [official OpenStack documentation](https://docs.openstack.org/ironic/7.0.1/api/ironic.common.states.html). ### Node lifecycle ![node-lifecyle](/images/hosted-private-cloud/opcp/node-lifecycle/03-node-lifecycle.png) When a node is installed and booted in an OPCP rack, its discovery is automatically performed by the control plane. At this moment, the node retrieves its properties and **traits** based on its hardware profile. Once the node is in the `Enroll` state, you can change its state so that it is managed by Ironic. **From the Horizon interface:** ![server-manageable](/images/hosted-private-cloud/opcp/node-lifecycle/03-server-status-to-manageable.png) **From the OpenStack APIs:** ```bash baremetal node manage $BAREMETAL_NODE_ID ``` To make the node available for installation, it must then be transitioned to the `Available` state: **From the Horizon interface:** ![server-available](/images/hosted-private-cloud/opcp/node-lifecycle/03-server-status-to-available.png) **From the OpenStack APIs:** ```bash baremetal node provide $BAREMETAL_NODE_ID ``` The node then transitions to the `Cleaning` state before reaching `Available`, making it deployable by the various projects in your OPCP environment. ### Maintenance mode This mode can be enabled to ensure a node cannot be used for installation, even if it is in the `Available` state. **From the Horizon interface:** ![Maintenance-on](/images/hosted-private-cloud/opcp/node-lifecycle/04-server-to-maintenance-on.png) When enabling maintenance, you may specify a reason so that the team responsible for the nodes has the information. This reason is optional. ![Maintenance-reason](/images/hosted-private-cloud/opcp/node-lifecycle/04-server-to-maintenance-reason.png) Once your maintenance operations are complete, you can disable maintenance: ![Maintenance-off](/images/hosted-private-cloud/opcp/node-lifecycle/04-server-to-maintenance-off.png) **From the OpenStack APIs:** ```bash baremetal node maintenance set $BAREMETAL_NODE_ID --reason "Maintenance reason" ``` You can then retrieve the maintenance status and the reason using the following command: ```bash baremetal node show $BAREMETAL_NODE_ID ``` You will find the following lines: ```bash | maintenance | True | maintenance_reason | "Maintenance reason" ``` To remove the node from maintenance, use the command: ```bash baremetal node maintenance unset $BAREMETAL_NODE_ID ``` ### References - [OpenStack Official Documentation - Horizon](https://docs.openstack.org/horizon/latest/) - [OpenStack Ironic States](https://docs.openstack.org/ironic/7.0.1/api/ironic.common.states.html) - [OpenStack Ironic Troubleshooting - Maintenance](https://docs.openstack.org/ironic/latest/install/troubleshooting.html) ## Go further If you need training or technical assistance for the implementation of our solutions, contact your sales representative or click [this link](https://www.ovhcloud.com/en-gb/professional-services/) to request a quote and have your project analyzed by our Professional Services team experts. Join our [community of users](https://community.ovhcloud.com/community/en).