Restoring Kubernetes Clusters Provisioned Using the Tanzu Kubernetes Grid Integrated Edition

This topic describes how to use BOSH Backup and Restore (BBR) to restore the BOSH Director, VMware Tanzu Kubernetes Grid Integrated Edition (TKGI) control plane, and Kubernetes clusters.

Overview

In the event of a disaster, you might lose your environment’s VMs, disks, and your IaaS network and load balancer resources as well. You can re-create your environment, configured with your saved Tanzu Kubernetes Grid Integrated Edition Ops Manager Installation settings, using your BBR backup artifacts.

Before restoring using BBR:

• Review the requirements listed in Compatibility of Restore below.
• Complete all of the steps documented in Preparing to Restore a Backup below.

Use BBR to restore the following:

• The BOSH Director plane, see Restore the BOSH Director below.
• The Tanzu Kubernetes Grid Integrated Edition control plane, see Restore Tanzu Kubernetes Grid Integrated Edition Control Plane below.
• The Tanzu Kubernetes Grid Integrated Edition clusters, see Restore Tanzu Kubernetes Grid Integrated Edition Clusters below.

Compatibility of Restore

The following are the requirements for a backup artifact to be restorable to another environment:

• Topology: BBR requires the BOSH topology of a deployment to be the same in the restore environment as it was in the back up environment.
• Naming of instance groups and jobs: For any deployment that implements the back up and restore scripts, the instance groups and jobs must have the same names.
• Number of instance groups and jobs: For instance groups and jobs that have back up and restore scripts, the same number of instances must exist.

Additional considerations:

• Limited validation: BBR puts the backed up data into the corresponding instance groups and jobs in the restored environment, but cannot validate the restore beyond that.
• Same Cluster: Currently, BBR supports the in-place restore of a cluster backup artifact onto the same cluster. Migration from one cluster to another using a BBR backup artifact has not yet been validated.

Note: This section is for guidance only. Always validate your backups by using the backup artifacts in a restore.

Prepare to Restore a Backup

Before you use BBR to either back up TKGI or restore TKGI from backup, follow these steps to retrieve deployment information and credentials:

• Verify your BBR Version
• Retrieve the BBR SSH Credentials
• Retrieve the BOSH Director Credentials
• Retrieve the UAA Client Credentials
• Retrieve the BOSH Director Address
• Download the Root CA Certificate
• Retrieve the BOSH Command Line Credentials
• Retrieve Your Cluster Deployment Names

Verify Your BBR Version

Before running BBR, verify that the installed version of BBR is compatible with the version of Ops Manager your TKGI tile is on:

1. To determine the Ops Manager BBR version requirements, see the Ops Manager Release Notes for the version of Ops Manager you are using.

2. To verify the currently installed BBR version, run the following command:

   bbr version  
   

3. If the installed BBR version does not meet the Ops Manager BBR version requirement, or BBR is not installed, you must upgrade BBR. For more information, see Installing BOSH Backup and Restore.

Retrieve the BBR SSH Credentials

There are two ways to retrieve BOSH Director credentials:

• Ops Manager Installation Dashboard
• Ops Manager API

Ops Manager Installation Dashboard

To retrieve your BBR SSH Credentials using the Ops Manager Installation Dashboard:

1. Navigate to the Ops Manager Installation Dashboard.
2. Click the BOSH Director tile.

3. Click the Credentials tab.

4. Locate Bbr Ssh Credentials.

5. Click Link to Credentials next to it.
6. Copy the private_key_pem field value.

Ops Manager API

To retrieve your BBR SSH Credentials using the Ops Manager API:

1. Obtain your UAA access token. For more information, see Access the Ops Manager API.

2. Retrieve the Bbr Ssh Credentials by running the following command:

   curl "https://OPS-MAN-FQDN/api/v0/deployed/director/credentials/bbr_ssh_credentials" \
   -X GET \
   -H "Authorization: Bearer UAA-ACCESS-TOKEN"
   

   Where:

   • OPS-MAN-FQDN is the fully-qualified domain name (FQDN) for your Ops Manager deployment.
   • UAA-ACCESS-TOKEN is your UAA access token.

3. Copy the value of the private_key_pem field.

Save the BBR SSH Credentials to File

To save the BBR SSH credentials to a private key file:

1. To reformat the copied private_key_pem value and save it to a file in the current directory:

   printf -- "YOUR-PRIVATE-KEY" > PRIVATE-KEY-FILE
   

   Where:

   • YOUR-PRIVATE-KEY is the text of your private key.
   • PRIVATE-KEY-FILE is the path to the private key file you are creating.

   For example:

   $ printf --  "-----begin rsa private key----- fake key contents ----end rsa private key-----" > bbr_key.pem
   

Retrieve the BOSH Director Credentials

There are two ways to retrieve BOSH Director credentials:

• Ops Manager Installation Dashboard
• Ops Manager API

Ops Manager Installation Dashboard

To retrieve your BOSH Director credentials using the Ops Manager Installation Dashboard, perform the following steps:

1. Navigate to the Ops Manager Installation Dashboard.
2. Click the BOSH Director tile.

3. Click the Credentials tab.

4. Locate Director Credentials.

5. Click Link to Credentials next to it.
6. Copy and record the value of the password field.

Ops Manager API

To retrieve your BOSH Director credentials using the Ops Manager API, perform the following steps:

1. Obtain your UAA access token. For more information, see Access the Ops Manager API.

2. Retrieve the Director Credentials by running the following command:

   curl "https://OPS-MAN-FQDN/api/v0/deployed/director/credentials/bbr_ssh_credentials" \
   -X GET \
   -H "Authorization: Bearer UAA-ACCESS-TOKEN"
   

   Where:

   • OPS-MAN-FQDN is the fully-qualified domain name (FQDN) for your Ops Manager deployment.
   • UAA-ACCESS-TOKEN is your UAA access token.

3. Copy and record the value of the password field.

Retrieve the UAA Client Credentials

To obtain BOSH credentials for your BBR operations, perform the following steps:

1. From the Ops Manager Installation Dashboard, click the Tanzu Kubernetes Grid Integrated Edition tile.
2. Select the Credentials tab.
3. Navigate to Credentials > UAA Client Credentials.
4. Record the value for uaa_client_secret.
5. Record the value for uaa_client_name.

Note: You must use BOSH credentials that limit the scope of BBR activity to your cluster deployments.

Retrieve the BOSH Director Address

You access the BOSH Director using an IP address.

To obtain your BOSH Director’s IP address:

1. Open the Ops Manager Installation Dashboard.
2. Select BOSH Director > Status.
3. Select the listed Director IP Address.

Log In To BOSH Director

1. If you are not using the Ops Manager VM as your jump box, install the latest BOSH CLI on your jump box.

2. To log in to BOSH Director, using the IP address that you recorded above, run the following command line:

   bosh -e BOSH-DIRECTOR-IP \
   --ca-cert PATH-TO-BOSH-SERVER-CERTIFICATE log-in
   

   Where:

   • BOSH-DIRECTOR-IP is the BOSH Director IP address recorded above.
   • PATH-TO-BOSH-SERVER-CERTIFICATE is the path to the root Certificate Authority (CA) certificate as outlined in Download the Root CA Certificate.

3. To specify Email, specify director.

4. To specify Password, enter the Director Credentials that you obtained in Retrieve the BOSH Director Credentials.
   
   For example:

   $ bosh -e 10.0.0.3 \
   --ca-cert /var/tempest/workspaces/default/root_ca_certificate log-in
   Email (): director
   Password (): *******************
   Successfully authenticated with UAA
   Succeeded
   

Download the Root CA Certificate

To download the root CA certificate for your Tanzu Kubernetes Grid Integrated Edition deployment, perform the following steps:

1. Open the Ops Manager Installation Dashboard.
2. In the top right corner, click your user name.
3. Navigate to Settings > Advanced.
4. Click Download Root CA Cert.

Retrieve the BOSH Command Line Credentials

1. Open the Ops Manager Installation Dashboard.
2. Click the BOSH Director tile.
3. In the BOSH Director tile, click the Credentials tab.
4. Navigate to Bosh Commandline Credentials.
5. Click Link to Credential.
6. Copy the credential value.

Retrieve Your Cluster Deployment Names

To locate and record a cluster deployment name, follow the steps below for each cluster:

1. On the command line, run the following command to log in:

   tkgi login -a TKGI-API -u USERNAME -k
   

   Where:

   • TKGI-API is the domain name for the TKGI API that you entered in Ops Manager > Tanzu Kubernetes Grid Integrated Edition > TKGI API > API Hostname (FQDN). For example, api.tkgi.example.com.
   • USERNAME is your user name.
     
     See Logging in to Tanzu Kubernetes Grid Integrated Edition for more information about the tkgi login command.

     Note: If your operator has configured Tanzu Kubernetes Grid Integrated Edition to use a SAML identity provider, you must include an additional SSO flag to use the above command. For information about the SSO flags, see the section for the above command in TKGI CLI. For information about configuring SAML, see Connecting Tanzu Kubernetes Grid Integrated Edition to a SAML Identity Provider

2. Identify the cluster ID:

   tkgi cluster CLUSTER-NAME
   

   Where CLUSTER-NAME is the name of your cluster.

3. From the output of this command, record the UUID value.

4. Open the Ops Manager Installation Dashboard.

5. Click the BOSH Director tile.

6. Select the Credentials tab.

7. Navigate to Bosh Commandline Credentials and click Link to Credential.

8. Copy the credential value.

9. SSH into your jump box. For more information about the jump box, see Installing BOSH Backup and Restore.

10. To retrieve your cluster deployment name, run the following command:

    BOSH-CLI-CREDENTIALS deployments | grep UUID
    

    Where:

    • BOSH-CLI-CREDENTIALS is the full value that you copied from the BOSH Director tile in Retrieve the BOSH Command Line Credentials.
    • UUID is the cluster UUID that you recorded in the previous step.
      
      

Transfer Artifacts to Your Jump Box

To restore BOSH director, Tanzu Kubernetes Grid Integrated Edition control plane or cluster you must transfer your BBR backup artifacts from your safe storage location to your jump box.

1. To copy an artifact onto a jump box, run the following SCP command:

   scp -r LOCAL-PATH-TO-BACKUP-ARTIFACT JUMP-BOX-USER@JUMP-BOX-ADDRESS:
   

   Where:

   • LOCAL-PATH-TO-BACKUP-ARTIFACT is the path to your BBR backup artifact.
   • JUMP-BOX-USER is the SSH user name of the jump box.
   • JUMP-BOX-ADDRESS is the IP address, or hostname, of the jump box.

2. (Optional) Decrypt your backup artifact if the artifact is encrypted.

Restore Kubernetes Clusters Provisioned by TKGI

Restoration of Kubernetes clusters provisioned by TKGI is a two step process: redeploy the clusters, then restore them.

Perform the following steps to redeploy the TKGI-provisioned Kubernetes clusters and restore their state from backup.

Redeploy Clusters

Before restoring your TKGI-provisioned clusters, you must redeploy them to BOSH. To redeploy TKGI-provisioned clusters:

• If you want to redeploy all clusters simultaneously, see Redeploy All Clusters.
• If you want to redeploy one cluster at a time, see Redeploy a Single Cluster.

Redeploy All Kubernetes Clusters

To redeploy all clusters:

1. In Ops Manager, navigate to the Tanzu Kubernetes Grid Integrated Edition tile.
2. Click Errands.
3. Ensure the Upgrade all clusters errand is On. This errand redeploys all your TKGI-provisioned clusters.
4. Return to the Installation Dashboard.
5. Click Review Pending Changes, review your changes, and then click Apply Changes. For more information, see Reviewing Pending Product Changes.

Redeploy a Single Kubernetes Cluster

To redeploy a TKGI-provisioned cluster through the TKGI CLI:

1. Identify the names of your TKGI-provisioned clusters:

   tkgi clusters
   

2. For each cluster you want to redeploy, run the following command:

   tkgi upgrade-cluster CLUSTER-NAME
   

   Where CLUSTER-NAME is the name of your Kubernetes cluster. For more information, see Upgrade Clusters.

Restore Kubernetes Clusters

After redeploying your TKGI-provisioned clusters, restore their stateless workloads and cluster state from backup by running the BOSH restore command from your jump box. Stateless workloads are tracked in the cluster etcd database, which BBR backs up.

Warning: BBR does not back up persistent volumes, load balancers, or other IaaS resources.

Warning: When you restore a cluster, etcd is stopped in the API server. During this process, only currently-deployed clusters function, and you cannot create new workloads.

To restore a cluster:

1. Move the cluster backup artifact to a folder from which you will run the BBR restore process.

2. SSH into your jump box. For more information about the jump box, see Configure Your Jump Box in Installing BOSH Backup and Restore.

3. Run the following command:

   BOSH_CLIENT_SECRET=BOSH-CLIENT-SECRET \
   nohup bbr deployment  --target BOSH-TARGET \
   --username BOSH-CLIENT  --deployment DEPLOYMENT-NAME \
   --ca-cert PATH-TO-BOSH-SERVER-CERT \
   restore \
   --artifact-path PATH-TO-DEPLOYMENT-BACKUP
   

   Where:

   • BOSH-CLIENT-SECRET is the BOSH_CLIENT_SECRET property. This value is in the BOSH Director tile under Credentials > Bosh Commandline Credentials.
   • BOSH-TARGET is the BOSH_ENVIRONMENT property. This value is in the BOSH Director tile under Credentials > Bosh Commandline Credentials. You must be able to reach the target address from the workstation where you run bbr commands.
   • BOSH-CLIENT is the BOSH_CLIENT property. This value is in the BOSH Director tile under Credentials > Bosh Commandline Credentials.
   • DEPLOYMENT-NAME is the cluster BOSH deployment name that you recorded in Retrieve Your Cluster Deployment Names above.
   • PATH-TO-BOSH-CA-CERT is the path to the root CA certificate that you downloaded in the Download the Root CA Certificate section above.
   • PATH-TO-DEPLOYMENT-BACKUP is the path to to your deployment backup. Make sure you have transfer your artifact into your jump box as described in Transfer Artifacts to Jump Box above.

   For example:

   $ BOSH_CLIENT_SECRET=p455w0rd \
   nohup bbr deployment \
   --target bosh.example.com \
   --username admin \
   --deployment service-instance_3839394 \
   --ca-cert bosh.ca.cert \
   restore \
   --artifact-path deployment-backup
   

   Note: The BBR restore command can take a long time to complete. The BBR restore command above uses nohup and the restore process is run within your SSH session. If you instead run the BBR command in a screen or tmux session the task will run separately from your SSH session and will continue to run, even if your SSH connection to the jump box fails.

4. To cancel a running bbr restore, see Cancel a Restore below.
5. After you restore a Kubernetes cluster, you must register its workers with their control plane nodes by following the Register Restored Worker VMs steps below.

6. If your Tanzu Kubernetes Grid Integrated Edition cluster restore fails, do one or more of the following:

   • Run the command again, adding the --debug flag to activate debug logs. For more information, see BBR Logging.
   • Follow the steps in Resolve a Failing BBR Restore Command below.

   Be sure to complete the steps in Clean Up After a Failed Restore below.

Register Restored Worker VMs

After restoring a Kubernetes cluster, you must register all of the cluster’s worker nodes with their control plane nodes. To register cluster worker nodes, complete the following:

1. Delete Nodes
2. Restart kubelet

Delete Nodes

To delete a cluster’s restored nodes:

1. To determine your cluster’s namespace, run the following command:

   kubectl get all --all-namespaces
   

2. To retrieve the list of worker nodes in the cluster, run the following command:

   kubectl get nodes -o wide
   

   Document the worker node names listed in the NAME column. Confirm the worker nodes are all listed with a status of NotReady.

3. To delete a node, run the following:

   kubectl delete node NODE-NAME
   

   Where NODE-NAME is a node NAME returned by the kubectl get nodes command.

4. Repeat the preceding kubectl delete node step for each of your cluster’s nodes.

Restart kubelet

To restart kubelet on your worker node VMs:

1. To restart kubelet on all of your cluster’s worker node VMs, run the following command:

   bosh ssh -d DEPLOYMENT-NAME worker -c 'sudo /var/vcap/bosh/bin/monit restart kubelet'
   

   Where DEPLOYMENT-NAME is the cluster BOSH deployment name that you recorded in Retrieve Your Cluster Deployment Names above.

2. To confirm all worker nodes in your cluster have been restored to a Ready state, run the following command:

   kubectl get nodes -o wide
   

Resolve a Failing BBR Restore Command

To resolve a failing BBR restore command:

1. Ensure that you set all the parameters in the command.
2. Ensure that the BOSH Director credentials are valid.
3. Ensure that the specified BOSH deployment or Director exists.
4. Ensure that the jump box can reach the BOSH Director.
5. Ensure the source backup artifact is compatible with the target BOSH deployment or Director.
6. If you see the error message Directory /var/vcap/store/bbr-backup already exists on instance, run the relevant commands from the Clean up After Failed Restore section of this topic.
7. See the BBR Logging topic.

Cancel a Restore

If you must cancel a restore, perform the following steps:

1. Terminate the BBR process by pressing Ctrl-C and typing yes to confirm.
2. Perform the procedures in the Clean up After Failed Restore section to support future restores. Stopping a restore can leave the system in an unusable state and prevent future restores.

Clean Up After a Failed Restore

If a BBR restore process fails, BBR might not have run the post-restore scripts, potentially leaving the instance in a locked state. Additionally, the BBR restore folder might remain on the target instance and subsequent restore attempts might also fail.

• To resolve issues following a failed BOSH Director restore, run the following BBR command:

  nohup bbr director \
  --host BOSH-DIRECTOR-IP \
  --username bbr \
  --private-key-path PRIVATE-KEY-FILE \
  restore-cleanup
  

  Where:

  • BOSH-DIRECTOR-IP is the address of the BOSH Director. If the BOSH Director is public, BOSH-DIRECTOR-IP is a URL, such as https://my-bosh.xxx.cf-app.com. Otherwise, this is the internal IP BOSH-DIRECTOR-IP which you can retrieve as show in Retrieve the BOSH Director Address above.
  • PRIVATE-KEY-FILE is the path to the private key file that you can create from Bbr Ssh Credentials as shown in Download the BBR SSH Credentials above.
    For example:

  $ nohup bbr director \
  --target 10.0.0.5 \
  --username bbr \
  --private-key-path private.pem \
  restore-cleanup
  

• To resolve issues following a failed control plane restore, run the following BBR command:

  BOSH_CLIENT_SECRET=BOSH-CLIENT-SECRET \
  bbr deployment \
  --target BOSH-TARGET \
  --username BOSH-CLIENT \
  --deployment DEPLOYMENT-NAME \
  --ca-cert PATH-TO-BOSH-CA-CERT \
  restore-cleanup
  

  Where:

  • BOSH-CLIENT-SECRET is the value for BOSH_CLIENT_SECRET retrieved in Download the BOSH Commandline Credentials above.
  • BOSH-TARGET is the value for BOSH_ENVIRONMENT retrieved in Download the BOSH Commandline Credentials above. You must be able to reach the target address from the workstation where you run bbr commands.
  • BOSH-CLIENT is the value for BOSH_CLIENT retrieved in Download the BOSH Commandline Credentials above.
  • DEPLOYMENT-NAME is the name retrieved in Retrieve Your Cluster Deployment Name above.
  • PATH-TO-BOSH-CA-CERT is the path to the root CA certificate that you downloaded in Download the Root CA Certificate above.
    For example:

  $ BOSH_CLIENT_SECRET=p455w0rd \
  bbr deployment \
  --target bosh.example.com \
  --username admin \
  --deployment pivotal-container-service-453f2f \
  --ca-cert bosh.ca.crt \
  restore-cleanup
  

• To resolve issues following a failed cluster restore, run the following BBR command:

  BOSH_CLIENT_SECRET=BOSH-CLIENT-SECRET \
  bbr deployment \
  --target BOSH-TARGET \
  --username BOSH-CLIENT \
  --deployment DEPLOYMENT-NAME \
  --ca-cert PATH-TO-BOSH-CA-CERT \
  restore-cleanup
  

  Where:

  • BOSH-CLIENT-SECRET is the value for BOSH_CLIENT_SECRET retrieved in Download the BOSH Commandline Credentials.
  • BOSH-TARGET is the value for BOSH_ENVIRONMENT retrieved in Download the BOSH Commandline Credentials. You must be able to reach the target address from the workstation where you run bbr commands.
  • BOSH-CLIENT is the value for BOSH_CLIENT retrieved in Download the BOSH Commandline Credentials.
  • DEPLOYMENT-NAME is the name retrieved in Retrieve Your Cluster Deployment Names above.
  • PATH-TO-BOSH-CA-CERT is the path to the root CA certificate that you downloaded in Download the Root CA Certificate.
    For example:

  $ BOSH_CLIENT_SECRET=p455w0rd \
  bbr deployment \
  --target bosh.example.com \
  --username admin \
  --deployment pivotal-container-service-453f2f \
  --ca-cert bosh.ca.crt \
  restore-cleanup