Newer documentation is now live.You are currently reading an older version.

Switch to latest✕

Delete a compute node

Since MOSK 23.2, the OpenStack-related metadata is automatically removed during the graceful machine deletion through the MOSK management console. For the procedure, refer to Delete a cluster machine.

During the graceful machine deletion, the OpenStack Controller (Rockoon) performs the following operations:

• Disables the OpenStack Compute and Block Storage services on the node to prevent further scheduling of workloads to it.

• Verifies if any resources are present on the node, for example, instances and volumes. By default, the OpenStack Controller blocks the removal process until the resources are removed by the user. To adjust this behavior to the needs of your cluster, refer to OpenStack Controller configuration.

• Removes OpenStack services metadata including compute services, Neutron agents, and volume services.

Caution

You cannot collocate the OpenStack compute node with other cluster components, such as Ceph. If done so, refer to the removal steps of the collocated components when planning the maintenance window.

If your cluster runs MOSK 23.1 or older version, perfrom the following steps before you remove the node from the cluster through the web UI to correctly remove the OpenStack-related metadata from it:

1. Disable the compute service to prevent spawning of new instances. In the keystone-client pod, run:

   openstack compute service set --disable <cmp_host_name> nova-compute --disable-reason "Compute is going to be removed."
   

2. Migrate all workloads from the node. For more information, follow Nova official documentation: Migrate instances.

3. Ensure that there are no pods running on the node to delete by draining the node as instructed in the Kubernetes official documentation: Safely drain node.

4. Delete the compute service using the OpenStack API. In the keystone-client pod, run:

   openstack compute service delete <service_id>
   

   Note

   To obtain <service_id>, run:

   openstack compute service list --host <cmp_host_name>
   

5. Depending on the networking backend in use, proceed with one of the following:

   Open vSwitch

   1. Obtain the network agent ID:

      openstack network agent list --host <cmp_host_name>
      

   2. Delete the Neutron Agent service running the following command in the keystone-client pod:

      openstack network agent delete <agent_id>
      

   Tungsten Fabric

   1. Log in to the Tungsten Fabric web UI.

   2. Navigate to Configure > Infrastructure > Virtual Routers.

   3. Select the target compute node.

   4. Click Delete.

6. Strongly recommended. Back up MKE as described in Mirantis Kubernetes Engine documentation: Back up MKE.

   Since the procedure above modifies the cluster configuration, a fresh backup is required to restore the cluster in case further reconfigurations fail.

   Important

   Because the MKE restoration process is complicated, we strongly recommend contacting Mirantis support for assistance.

   If you still decide to restore MKE from a backup on your own, you must scale down helm-controller on the cluster being restored if the MKE version of the affected cluster after the restore will differ from the MKE version in the ClusterRelease object that is set in MOSK Cluster objects in the management cluster:

   • If you are restoring MKE on a management cluster: before starting the restore, scale down helm-controller on each affected MOSK cluster. This prevents unintended Ceph and OpenStack downgrades on MOSK clusters after the management cluster is restored.

   • If you are restoring MKE on a MOSK cluster: immediately after the restore completes, scale down helm-controller. Because the restore rolls the cluster back to an older release, this prevents it from triggering a premature upgrade of Helm releases.