User guide

compute nodes, you purchase what are called reserved node offerings.You purchase one offering for
each compute node that you want to reserve.When you reserve a compute node, you pay a fixed up-front
charge and then an hourly recurring charge, whether your cluster is running or not. The hourly charges,
however, are significantly lower than those for on-demand usage. For more information, see Purchasing
Amazon Redshift Reserved Nodes (p. 107).
Resizing a Cluster
If your storage and performance needs change after you initially provision your cluster, you can resize
your cluster.You can scale the cluster in or out by adding or removing nodes. Additionally, you can scale
the cluster up or down by specifying a different node type.
For example, you can add more nodes, change node types, change a single-node cluster to a multi-node
cluster, or change a multi-node cluster to a single-node cluster.You must, however, ensure that the
resulting cluster is large enough to hold the data that you currently have or else the resize will fail. When
using the API, you have to specify both the node type and the number of nodes even if you only change
one of the two.
When you resize a cluster, Amazon Redshift first puts your existing cluster into read-only mode. Amazon
Redshift then provisions the nodes as requested, copies data to the new cluster, and switches your
connections to use the new cluster. At this time, you lose any open connections to the old cluster. If you
have any queries in progress at the time of the switch, you will notice the connection loss.You must
restart the query on the new cluster.You should wait until the resize operation is complete before you
resume loading data or performing queries on the cluster.
When you resize your cluster, it will remain in read-only mode until the resize finishes.You can view the
resize progress on the cluster's Status tab in the Amazon Redshift console.The time it takes to resize
a cluster depends on the amount of data in each node.Typically, the resize process varies from a couple
of hours to a day, although clusters with larger amounts of data might take even longer. This is because
the data is copied in parallel from each node on the source cluster to the nodes in the target cluster. For
more information about resizing clusters, see Tutorial: Resizing Clusters in Amazon Redshift (p.224) and
Resizing a Cluster (p. 24).
Amazon Redshift does not sort tables during a resize operation. When you resize a cluster, Amazon
Redshift distributes the database tables to the new compute nodes based on their distribution styles and
runs an ANALYZE to update statistics. Rows that are marked for deletion are not transferred, so you will
only need to run a VACUUM if your tables need to be resorted. For more information, see Vacuuming
tables in the Amazon Redshift Database Developer Guide.
If your cluster is public and is in a VPC, it keeps the same elastic IP address (EIP) for the leader node
after resizing. If your cluster is private and is in a VPC, it keeps the same private IP address for the leader
node after resizing. If your cluster is not in a VPC, a new public IP address is assigned for the leader
node as part of the resize operation.
To get the leader node IP address for a cluster, use the dig utility, as shown following:
dig mycluster.abcd1234.us-west-2.redshift.amazonaws.com
The leader node IP address is at the end of the ANSWER SECTION in the results, as shown following:
API Version 2012-12-01
8
Amazon Redshift Management Guide
Resizing a Cluster