Releasing nodes from the compute cluster

If you need to release compute nodes, you can start with regular (non-management) nodes. When all of the regular compute nodes are released, you can proceed to remove the management nodes. The management nodes can only be released together. Their release destroys the compute cluster.

Limitations

The compute cluster must have at least three nodes, to allow self-service users to enable high availability for Kubernetes master nodes.

Prerequisites

A clear understanding of the limitations listed in High availability and the compute cluster.
If the node hosts virtual machines, they should be migrated to other nodes, as described in Migrating virtual machines.
To destroy the compute cluster, all virtual machines must be deleted.

To release nodes from the compute cluster

Admin panel

On the Compute > Nodes screen, do one of the following:
- Select the nodes, and then click Release nodes above the list.
- Click the ellipsis icon next to a node and select Release.
- Click a node to open its details, and then click Release node on the node right pane.
In the Release node window, confirm the action by clicking Release.

The selected nodes will disappear from the Nodes screen. If you selected all of the compute nodes, the compute cluster will be destroyed.

Command-line interface

Use the following command:

vinfra service compute node release [--compute] [--controller] <node>

--compute: Compute node role
--controller: Compute controller node role
<node>: Node ID or hostname

For example, to release the node node005.vstoragedomain from the compute cluster, run:

# vinfra service compute node release node005.vstoragedomain

The removed node will disappear from the vinfra service compute node list output:

# vinfra service compute node list
+------------------+------------------------+---------+--------------+
| id               | host                   | state   | roles        |
+------------------+------------------------+---------+--------------+
| 7ffa9540-5a20<…> | node001.vstoragedomain | healthy | - controller |
|                  |                        |         | - compute    |
| 6e8afc28-7f71<…> | node002.vstoragedomain | healthy | - compute    |
| 02ff64ae-5800<…> | node003.vstoragedomain | healthy | - compute    |
| 827a1f4e-56e5<…> | node004.vstoragedomain | healthy | - compute    |
+------------------+------------------------+---------+--------------+

To release all nodes from the compute cluster, use the following command:

vinfra service compute delete