Anthos clusters on bare metal is now Google Distributed Cloud (software only) for bare metal. For more information, see the product overview.

Reset nodes and delete clusters

When a Google Distributed Cloud cluster is in the process of being installed, binaries and systemd services are installed on the nodes hosting that cluster, and services begin to listen to ports on the nodes.

However, if a cluster installation fails, all these binaries and services need to be deleted. In other words, the nodes need to be reset or wiped clean to prepare them for a reattempt at installing the cluster. If nodes aren't reset in this way, the next attempt to install a cluster on them fails.

This page describes how do this clean up operation of specific nodes and how to delete a cluster.

This page is for Admins and architects and Operators who manage the lifecycle of the underlying tech infrastructure. To learn more about common roles and example tasks that we reference in Google Cloud content, see Common GKE Enterprise user roles and tasks.

Choose a deletion method

The method that you use to delete a cluster depends on:

The cluster type.
If you want to clean up only specific nodes and not delete the entire cluster.
How the cluster was created.

Google Distributed Cloud provides the following deletion methods:

The Google Cloud console or Google Cloud CLI:
- Use the console or gcloud CLI to delete user clusters that are managed by the GKE On-Prem API. A user cluster is managed by the GKE On-Prem API if one of the following is true:
  - The cluster was created in the Google Cloud console or using the gcloud CLI, which automatically configures the GKE On-Prem API to manage the cluster.
  - The cluster was created using bmctl, but it was configured it to be managed by the GKE On-Prem API.
bmctl:
- Use bmctl reset nodes to reset specific nodes.
- Use bmctl reset to delete the following cluster types:
  - Admin, hybrid, and standalone clusters (referred to as self-managed clusters). This includes admin clusters that are managed by the GKE On-Prem API.
  - User clusters that aren't managed by the GKE On-Prem API.
If you use bmctl to reset nodes or to delete a cluster, the command expects the cluster configuration file to be in the current working directory. By default, the path is like the following:

bmctl-workspace/CLUSTER_NAME/CLUSTER_NAME.yaml

If you used the --workspace-dir flag to specify a different directory during cluster creation, you must use the flag to specify the working directory during cluster reset.
kubectl:
- Use kubectl delete cluster to delete only user clusters that aren't managed by the GKE On-Prem API clusters. Don't run the command on other cluster types.
- Note that if you use kubectl delete cluster, you must also delete the namespace that the cluster is in after you delete the cluster.

After you delete a cluster, you can reinstall it after making any needed configuration changes.

Delete self-managed clusters

To delete an admin, hybrid, or standalone cluster, run the following command:

bmctl reset --cluster CLUSTER_NAME

In the command, replace CLUSTER_NAME with the name of the cluster you want to reset.

Output from the bmctl cluster reset command looks similar to this sample:

Please check the logs at bmctl-workspace/example-cluster-1/log/reset-20221025-184705/reset.log
[2022-10-25 18:47:11+0000] Creating bootstrap cluster... OK
[2022-10-25 18:48:18+0000] Loading images... OK
[2022-10-25 18:48:18+0000] Waiting for reset jobs to finish...
[2022-10-25 18:48:28+0000] Operation reset in progress: 1       Completed: 0    Failed: 0
...
[2022-10-25 18:50:08+0000] Operation reset in progress: 0       Completed: 1    Failed: 0
[2022-10-25 18:50:08+0000] Flushing logs... OK
[2022-10-25 18:50:08+0000] Deleting GKE Hub member example-cluster-1 in project example-project-12345...
[2022-10-25 18:50:11+0000] Successfully deleted GKE Hub member example-cluster-1 in project example-project-12345
[2022-10-25 18:50:11+0000] Deleting bootstrap cluster... OK

In addition to deleting the cluster, the command deletes the cluster's membership from the fleet.

For admin clusters managed by the GKE On-Prem API, you also need to delete the API resources in Google Cloud. Otherwise, the cluster will be displayed on the GKE clusters page in the Google Cloud console. Use the following command to delete the GKE On-Prem API resources for an admin cluster:

gcloud container bare-metal admin-clusters unenroll CLUSTER_NAME \
    --project=FLEET_HOST_PROJECT_ID \
    --location=REGION \
    --ignore-errors

Replace the following:

FLEET_HOST_PROJECT_ID: The project ID of the fleet in which the admin cluster was a member.
REGION: The Google Cloud region in which the GKE On-Prem API stores cluster metadata.

The --ignore-errors flag ensures that the unenrollment of a bare metal admin cluster resource succeeds even if errors occur during unenrollment.

After the cluster deletion finishes, you can create a new cluster. For details, see Cluster creation overview.

Delete user clusters

If the user cluster is managed by the GKE On-Prem API, delete the cluster using the console or the gcloud CLI. Otherwise, use bmctl or kubectl to delete the cluster.

`bmctl`

You can use bmctl to delete user clusters that were created with bmctl or kubectl, and that aren't enrolled in the GKE On-Prem API.

Run the following command to delete a user cluster with bmctl:

bmctl reset --cluster USER_CLUSTER_NAME --admin-kubeconfig ADMIN_KUBECONFIG_PATH

In the command, replace the following entries with information specific to your environment:

USER_CLUSTER_NAME: the name of the user cluster you're deleting.
ADMIN_KUBECONFIG_PATH: the path to the associated admin cluster's kubeconfig file. bmctl supports the use of --kubeconfig as an alias for the--admin-kubeconfig flag.

Output from the bmctl cluster reset command looks similar to this sample:

Please check the logs at bmctl-workspace/example-cluster-1/log/reset-20221025-184705/reset.log
[2022-10-25 18:47:11+0000] Creating bootstrap cluster... OK
[2022-10-25 18:48:18+0000] Loading images... OK
[2022-10-25 18:48:18+0000] Waiting for reset jobs to finish...
[2022-10-25 18:48:28+0000] Operation reset in progress: 1       Completed: 0    Failed: 0
...
[2022-10-25 18:50:08+0000] Operation reset in progress: 0       Completed: 1    Failed: 0
[2022-10-25 18:50:08+0000] Flushing logs... OK
[2022-10-25 18:50:08+0000] Deleting GKE Hub member example-cluster-1 in project example-project-12345...
[2022-10-25 18:50:11+0000] Successfully deleted GKE Hub member example-cluster-1 in project example-project-12345
[2022-10-25 18:50:11+0000] Deleting bootstrap cluster... OK

`kubectl`

You can use kubectl to delete user clusters that were created with bmctl or kubectl, and that aren't enrolled in the GKE On-Prem API. To use kubectl to delete a user cluster, you must first delete the cluster object, then its namespace. Otherwise, the jobs to reset machines can't be created, and the deletion process might be stuck indefinitely.

To delete a user cluster with kubectl:

Run the following command to delete the cluster object:
```
kubectl delete cluster USER_CLUSTER_NAME -n USER_CLUSTER_NAMESPACE \
    --kubeconfig ADMIN_KUBECONFIG_PATH
```
In the command, replace the following entries with information specific to your environment:
- USER_CLUSTER_NAME: the name of the user cluster you're deleting.
- USER_CLUSTER_NAMESPACE: the namespace for the cluster. By default, the cluster namespaces for Google Distributed Cloud are the name of the cluster prefaced with cluster-. For example, if you name your cluster test, the namespace has a name like cluster-test.
- ADMIN_KUBECONFIG_PATH: the path to the associated admin cluster's kubeconfig file.
After the cluster is deleted successfully, run the following command to delete the namespace:
```
kubectl delete namespace USER_CLUSTER_NAMESPACE --kubeconfig ADMIN_KUBECONFIG_PATH
```

Console

If the user cluster is managed by the GKE On-Prem API do the following steps to delete the cluster:

In the console, go to the Google Kubernetes Engine clusters overview page.

Go to GKE clusters
Select the Google Cloud project that the user cluster is in.
In the list of clusters, click the cluster that you want to delete.
In the list of clusters, locate the cluster that you want to delete. If the Type is external, this indicates that the cluster was created using bmctl and wasn't enrolled in the GKE On-Prem API. In this case, follow the steps in the bmctl or kubectl tab to delete the cluster.

If the icon in the Status column indicates a problem, follow the steps in the gcloud CLI tab to delete the cluster. You will need to add the --ignore-errors flag to the delete command.
Click the name of the cluster that you want to delete.
In the Details panel, near the top of the window, click Delete.
When prompted to confirm, enter the name of the cluster and click Remove.

gcloud CLI

If the user cluster is managed by the GKE On-Prem API, do the following steps to delete the cluster on a computer that has the gcloud CLI installed:

Log in with your Google account:
```
gcloud auth login
```
Update components:
```
gcloud components update
```
Get a list of clusters to help ensure that you specify the correct cluster name in the delete command:
```
gcloud container bare-metal clusters list \
  --project=FLEET_HOST_PROJECT_ID \
  --location=LOCATION
```
Replace the following:
- FLEET_HOST_PROJECT_ID: The ID of the project that the cluster was created in.
- LOCATION: The Google Cloud location associated with the user cluster.
The output is similar to the following:
```
NAME                      LOCATION    VERSION         ADMIN_CLUSTER            STATE
example-user-cluster-1a   us-west1    1.31.0-gke.889          example-admin-cluster-1  RUNNING
```
Run the following command to delete the cluster:
```
gcloud container bare-metal clusters delete USER_CLUSTER_NAME \
  --project=FLEET_HOST_PROJECT_ID \
  --location=LOCATION \
  --force \
  --allow-missing
```
Replace the following:
- USER_CLUSTER_NAME: The name of the user cluster to delete.
- FLEET_HOST_PROJECT_ID: The ID of the project that the cluster was created in.
- LOCATION: The Google Cloud location associated with the user cluster.
The --force flag lets you delete a cluster that has node pools. Without the --force flag, you have to delete the node pools first, and then delete the cluster.

The --allow-missing flag is a standard Google API flag. When you include this flag, the command returns success if the cluster isn't found.

If the command returns an error that contains the text failed connecting to the cluster's control plane, this indicates connectivity issues with either the admin cluster, the Connect Agent, or the on-premises environment. To troubleshoot issues with the Connect Agent, see Collecting Connect Agent logs.
- If you think the connectivity issue is transient, for example, because of network problems, wait and retry the command.
- If you know that the admin cluster has been deleted, of if the node machines for the admin or the user cluster have been shut down or taken offline, include the --ignore-errors flag and retry the command.
  
  You also need to include --ignore-errors if the cluster was deleted using bmctl or kubectl, which leaves GKE On-Prem API resources in Google Cloud. One symptom of this is that the cluster is still displayed on the GKE clusters page in the console in a unhealthy state.
Caution: When the admin and user clusters are unreachable, and you include --ignore-errors, the delete command removes Google Cloud resources only. In this case, you still need to delete the cluster using bmctl.

For information about other flags, see the gcloud CLI reference.

Reset specific cluster nodes

You might want to reset specific nodes of a cluster if, for example, an admin cluster has been deleted but the user clusters managed by that admin cluster remain. In this case, the user clusters as a whole can't be deleted because the admin cluster has been deleted. Consequently, the nodes of the user clusters have to be individually reset.

To reset nodes, you need a service account with read access to Google Container Registry (GCR). The bmctl command expects the JSON key file for this service account as an argument. To reset individual nodes of a cluster, run the following command:

bmctl reset nodes --addresses NODE_1_IP_ADDRESS,NODE_2_IP_ADDRESS \
    --ssh-private-key-path SSH_KEY_PATH \
    --gcr-service-account-key SERVICE_ACCOUNT_KEY_PATH \
    --login-user root

In the command, replace the following entries with information specific to your environment:

NODE_1_IP_ADDRESS , NODE_2_IP_ADDRESS: comma-separated list of IP addresses of nodes you want to delete.
SSH_KEY_PATH: path to SSH private key. This is the key that will be used to establish SSH connections with nodes during reset.
SERVICE_ACCOUNT_KEY_PATH: path to the JSON file that contains the service account key. This key gives bmctl permission to pull images from the Google Container Registry. You can create a service account key using the console or the gcloud CLI. For details, see Creating and managing service account keys. Another way the service account key file can be created is if you run the create config command with the --create-service-accounts flag. For details about that command, see Create and admin cluster config with bmctl.

Note: Service account keys are a security risk if not managed correctly. You are responsible for the security of the private key and for other operations described by Best practices for managing service account keys. If you are prevented from creating a service account key, service account key creation might be disabled for your organization. For more information, see Managing secure-by-default organization resources.

Cluster deletion details

During the deletion, the cluster's fleet membership registration, storage mounts, and data from the anthos-system StorageClass are deleted.

For all nodes, the tunnel interfaces used for cluster networking are removed, and the following directories are deleted:

/etc/kubernetes
/etc/cni/net.d
/root/.kube
/var/lib/kubelet

For load balancer nodes:

The keepalived and haproxy services are deleted.
The configuration files for keepalived and haproxy are deleted.