Node cleaning

In Ironic cleaning is a process of preparing a bare metal node for provisioning. There are two types of cleaning: automated and manual. See cleaning documentation for more details.

Warning

It is highly recommended to at least wipe metadata (partitions and partition table(s)) from all disks before deployment.

Automated cleaning

Automated cleaning runs before a node gets to the available state (see Bare Metal Node States for more information on provisioning states). It happens after the first enrollment and after every unprovisioning.

In the TripleO undercloud automated cleaning is disabled by default. Starting with the Ocata release, it can be enabled by setting the following option in your undercloud.conf:

[DEFAULT]
clean_nodes = True

Alternatively, you can use Manual cleaning as described below.

Manual cleaning

Manual cleaning is run on request for nodes in the manageable state.

If you have automated cleaning disabled, you can use the following procedure to wipe the node’s metadata starting with the Rocky release:

  1. If the node is not in the manageable state, move it there:

    openstack baremetal node manage <UUID or name>
    
  2. Run manual cleaning on a specific node:

    openstack overcloud node clean <UUID or name>
    

    or all manageable nodes:

    openstack overcloud node clean --all-manageable
    
  3. Make the node available again:

    openstack overcloud node provide <UUID or name>
    

    or provide all manageable nodes:

    openstack overcloud node provide --all-manageable
    

See Cleaning Nodes for an explanation how to use Mistral workflows directly to initiate cleaning. This is particularly useful if you want to run some non-standard clean steps.