[Bug]: resource/aws_elasticache_replication_group: Prevents `automatic_failover_enabled` being off in cluster_mode

[aolley]

Terraform Core Version

1.5.7

AWS Provider Version

5.66.0

Affected Resource(s)

• aws_elasticache_replication_group

Expected Behavior

Setting a replication group with:

• cluster_mode = "enabled"
• automatic_failover_enabled = true
• num_node_groups = 1

Should be possible.

Actual Behavior

Terraform validation prevents it.

Relevant Error/Panic Output Snippet

Planning failed. Terraform encountered an error while generating this plan.

╷
│ Error: "num_cache_clusters": must be at least 2 if automatic_failover_enabled is true

Terraform Configuration Files

No specific config needed.

Steps to Reproduce

Try to plan a replication group with:

• cluster_mode = "enabled"
• automatic_failover_enabled = true
• num_node_groups = 1

Debug Output

No response

Panic Output

No response

Important Factoids

This was a problem in the past, it was fixed in this PR for 3.36: https://github.com/hashicorp/terraform-provider-aws/pull/18635

The same issue was (mistakenly?) re-introduced in 5.59: https://github.com/hashicorp/terraform-provider-aws/pull/38396

Note, AWS API returns the following if you try to turn off automatic failover:

InvalidParameterCombination: Turning off automatic failover for replication group with cluster mode enabled is not supported.

References

No response

Would you like to implement a fix?

None

[github-actions[bot]]

Community Note

Voting for Prioritization

• Please vote on this issue by adding a 👍 reaction to the original post to help the community and maintainers prioritize this request.
• Please see our prioritization guide for information on how we prioritize.
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request.

Volunteering to Work on This Issue

• If you are interested in working on this issue, please leave a comment.
• If this would be your first contribution, please review the contribution guide.

[aolley]

@gdavison pinging you since you likely have the most context here given your involvement in the previous PRs. Are we able to fine tune the validation to handle when cluster_mode is enabled?