Why can’t I perform an elastic resize for my Amazon Redshift cluster?

Resolution

Note: If you receive errors when you run AWS Command Line Interface (AWS CLI) commands, then see Troubleshooting errors for the AWS CLI. Also, make sure that you're using the most recent AWS CLI version.

You can use the elastic resize feature in Amazon Redshift to quickly add or remove nodes from your cluster, with minimal downtime. However, there are constraints for when you can use elastic resize that are based on factors such as the node type and original cluster size. These constraints might also cause a brief interruption to queries and connections.

Note: Elastic resize might not be available for data sharing clusters or when you transfer data from a shared snapshot. Elastic resize doesn't sort tables or reclaim disk space automatically.

In Amazon Redshift, slice shuffling happens during an elastic resize when the cluster's slice-to-node mapping changes. Because Amazon Redshift evenly distributes data across slices for parallel processing, slices are redistributed when you add or remove nodes. This redistribution balances data across the new slices to maintain efficient performance. The time it takes to perform the elastic resize depends on cluster size, data volume, and workload activity.

Before you update the node count for your Amazon Redshift cluster, use the DescribeNodeConfigurationOptions API to determine the appropriate node configurations for an elastic resize. For example, the node count and node type.

Example:

aws redshift describe-node-configuration-options --cluster-identifier cluster-name --region aws-region --action-type resize-cluster

Note: Replace cluster-name with your cluster name. Replace aws-region with your AWS Region.

The preceding example checks the available node configuration options for elastic resize.

Amazon Redshift elastic resize growth and reduction limits are different for each node type:

• For dc2.large node types, you can only double the node count or decrease the node count by half of the original cluster.
• For dc2.8xlarge node types, you can resize up to 2 times the original node count, or resize down to half of the original node count. For example, you can resize a 16-node cluster to any size that is between 8 and 32 nodes.
• For ra3.16xlarge or ra3.4xlarge node types, you can resize up to 4 times the original node count. You can also resize down to one quarter of the original node count. For example, you can resize a 16-node cluster to any size that is between 4 and 64 nodes.
• For ra3.xlplus node types, you can resize up to 2 times the original node count, or resize down to a quarter of the original node count. For example, you can resize a 16-node cluster to any size that is between 4 and 32 nodes.
• For ra3.xlarge node types, you can resize up to 2 times the original node count, or resize down to half of the original node count.

Note:

• The cluster must be in a virtual private cloud (VPC).
• For Amazon Redshift RA3 nodes, there might be limitations based on the amount of managed storage used.

Related information

How do I resize an Amazon Redshift cluster?

Cluster operations