Scaling up the target cluster

This step is applicable only when you are following the Side-by-side upgrade (single node start).

To scale up the target cluster, follow these steps:

1. Perform sanity check on the single node target cluster. Make sure it is functional and the migrated data is intact.

2. Check the Scaling a single-node (evaluation) deployment to a multi-node (HA) deployment page for the detailed steps on converting a single-node deployment to a multi-node setup.

   To reduce the infra requirements, you can clean the nodes from the source cluster and add it to the target cluster one by one. Do not dismantle all the nodes of source cluster at once.

   For removing the nodes from the source cluster, check the How to uninstall the cluster page. Before you proceed with the uninstallation of Kubernetes, ensure you delete the node from the cluster using kubectl delete node.

   The nodes set to join the target cluster must be clean and should not contain any data corresponding to the old cluster.

   Note: If you are removing the nodes from source cluster and joining to the target cluster, it is strongly recommended to create a backup of the source cluster and the SQL database before the scale up. This is to ensure you can restore the source cluster if something goes wrong while scaling up the target cluster.

3. After converting the target cluster to HA setup, you can retain the same FQDN as the source cluster or update the FQDN. After updating the FQDN, switch the traffic to the new cluster.

   • Before updating FQDN, you must add the required certificates to match the FQDN.
   • To update the FQDN, refer to the Configuring the FQDN post-installation page.
   Note:

   Changing the FQDN also requires new server certificates. If a new certificate is available, you have two options: either continue with the new self-signed certificates configured by the installer automatically, or stop the installation and bring in a new certificate.

   You can configure the certificate via the server_certificate field in cluster_config.json.

4. In case of failure or issues during the side-by-side upgrade process, you can instantly roll back to your previous state by disabling maintenance mode on your source cluster by following the Instant rollback on failure page.

   If you have decommissioned any nodes from source cluster, you can add them back by following the Adding a new node to the cluster page.

5. Enable the backup on the cluster target.