Automatic Cluster Creation

Accelerator needs to launch a cluster in order to hold your database. After you have successfully registered and completed the one-time cross account linking process, Accelerator will collect the required permissions by generating the CFT (5 IAM policy and 1 IAM Role) use the permissions to create the cloud stack (CloudFormation stack). Based on the configurations provided (number of nodes, type of EC2 instances, region, availability zone, etc.), Accelerator launches the EC2 instances. Network configuration (VPC, subnet etc.) will then be created with a specific naming pattern (included database name) and will be attached to the EC2 instances. Accelerator also uses CloudWatch rules and events to receive the status of the resource that Accelerator creates for you. If all goes well, a Vertica cluster is created.

Out of the Box Subcluster Configurations

The Quick Setup cluster configuration option allows you to choose from several template subcluster configurations. Each option corresponds to a different use case scenario: Extract, Transform, Load (ETL); Dashboard; and Query Workload. Either select one of these default templates, or choose the Advanced Setup option and choose your own custom options, as outlined below.

Set Your AWS Resources

AWS Account default resource limits are typically insufficient for creating clusters with more than a few nodes.

For larger clusters, or for busy AWS accounts, you will need:

Up to/at least 1 available VPC for each database you plan to create (limit: increase VPC, default: five per region). Ten VPCs are recommended if you plan to create more databases.
At least one Gateway VPC Endpoint per Database you plan to create (limit: Gateway VPC endpoints per region, default: 20 per region). A minimum limit that is equal to or greater than your VPC limit is recommended.
Up to 48 Elastic IPs for each database you plan to create (allowing for a maximum size of 48 nodes). (Limit: new VPC elastic IP address limit, default: five per region) This value depends on the number of nodes you select. Of course, each node needs an IP. It is recommend that you get 100 EIPs just in case you are creating large clusters or have a busy AWS account.
Up to 3072 vCPUs per cluster 48 x i3.16x Instance (limit: all standard (A, C, D, H, I, M, R, T, Z) instances, default: 200 vCPUs) The smallest node Accelerator uses has eight vCPUs, and larger nodes of bigger instances require this limit be increased based on your database size.

AWS Service Limits

You can increase your service limits by opening a support ticket with AWS. AWS typically takes a day to increase your service limits.

Please Note:

Limits are region-specific. (Please increase your service limits in all regions you plan to use)
If you are already using resources, you will need to make sure the resources are accounted for by increasing limits accordingly.

Warning: Create database or subcluster can fail if you do not have adequate resources.