Scaling Subclusters Up or Down in MC
You can scale an Eon Mode subcluster up or down, to increase or decrease the number of nodes in the subcluster. This lets you add compute capacity when you need it, and reduce it to save money or redirect resources when you don't.
For Eon Mode databases, MC supports actions for subcluster and node management:
- In the cloud: On AWS and GCP.
- On-premises: For Pure Storage FlashBlade.
In Enterprise Mode, subclusters are not supported.
Scaling Up or Down in the Cloud
Scaling Up a Subcluster
When you scale up a subcluster, MC adds one or more cloud instances to your database cluster as hosts, and adds them to your subcluster as nodes.
- On the Manage > Subclusters tab, click Scale Up immediately above the subcluster you want to enlarge. MC launches the Scale Up wizard.
- Wait a moment or two while MC pre-populates the fields on the Enter AWS Credentials and preferences screen with your credentials, then click Next.
-
On the Specify Subcluster information screen, enter the number of instances you want to add to the subcluster in Number of Instances to Add. MC pre-populates the number of existing hosts in the cluster,
- Click Browse and select your Vertica license to insert in the license field, then click Next.
-
The Specify cloud instance and data storage info screen has fields that specify the cloud instance and data storage information. MC sets all the fields to the same values as the existing instances in your subcluster. Accept the defaults by clicking Next.
- The Specify additional storage and tag info screen displays any tags you specified for your instances, when you created the database cluster. You can use the same tags for the instances you are adding, or use the fields on the screen to add new tags for these instances. You can keep and apply or delete the existing tags. If you delete tags, they are not used for the instances you are adding now. When you are done modifying the tags, click Next.
- MC displays the Review 'Scale Up' information screen for your approval. If the information is correct, click Scale Up to add the instances to the subcluster.
- Wait until all operations on the progress screen show check marks. Then click Close.
Scaling Down a Subcluster
When you scale down a subcluster, MC removes the requested number of nodes from the subcluster, and removes their hosts from the database cluster. It then terminates the underlying cloud instances. Scaling down a subcluster is allowed only when doing so will not cause the database to shut down.
For details on when Scale Down is or is not available for a subcluster, see Subcluster Action Rules in MC.
-
On the Manage > Subclusters tab, click Scale Down above the subcluster. MC launches the Scale Down wizard.
- In the Number of hosts to remove field, enter the number of hosts you would like to remove from the subcluster.
- Under Confirmation, click the checkbox to affirm that you want to scale down the named subcluster, and terminate the instances the nodes were using.
-
Click Scaledown Subcluster.
The scale down operation may take a little time. When it finishes, the progress screen displays the details about the removed nodes and their hosts and the terminated instances.
- When all steps show check marks, click Close.
Scaling Up or Down on Premises
When you scale up a subcluster on-premises, the MC wizard displays a list of the available Vertica host machines that are not currently part of a database. You select the ones you want to add to the subcluster as nodes, then confirm that you want to scale up the subcluster. When you scale down a subcluster on-premises, MC removes the nodes from the subcluster in the database, but does not terminate the Vertica host machines. The hosts are now available for scaling up other subclusters.