Aws_redshift_cluster can't be resized without changing number_of_nodes

[hashibot]

This issue was originally opened by @wojukasz as hashicorp/terraform#11402. It was migrated here as part of the provider split. The original body of the issue is below.

---

I am trying to updated redshift cluster from dc1.large to dc1.8xlarge while maintaining number of nodes (currently set at 2 and previously declared to this value) but I am getting an error that multi-node must be greater than or equal to 2?

Terraform Version : 0.8.4
Affected Resource: aws_redshift_cluster
Terraform Configuration:
(...)
cluster_type = "multi-node"
number_of_nodes = "2"
node_type = "dc1.8xlarge" #only change from dc1.large
(...)

Expected Behavior: Be able to change the instance/cluster size without modifying number of nodes in use.

Actual Behavior:
module.redshift_cluster.aws_redshift_cluster.cluster: Modifying...
node_type: "dc1.large" => "dc1.8xlarge"

Error applying plan:

1 error(s) occurred:

* aws_redshift_cluster.cluster: [WARN] Error modifying Redshift Cluster (redshift-cluster-dev): InvalidParameterCombination: Number of nodes for cluster type multi-node must be greater than or equal to 2
status code: 400, request id: 500f2ce2-e2d2-11e6-b32e-b73a2dabc4c7

Steps to Reproduce: Create dc1.large multi-node cluster with 2 nodes and change the cluster size dc1.8xlarge

Seems to be related to: hashicorp/terraform#6274

[foragerr]

absolutely related to hashicorp/terraform#6274 in that, both node type and node count need to be passed when updating only one of the two with modify-cluster.

[bflad]

The fix for this has been merged in via #3127 and will release with v1.18.0 of the AWS provider in two days.

[bflad]

This has been released in version 1.18.0 of the AWS provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!