1.13.0 is forcing a recreate of the cluster instance

[oycyc]

Describe the Bug

I'm currently using 1.12.0, which is good and all.

I'd like to upgrade to 1.14.0 which has a bug fix in an output that I'd like to utilize.

BUT a PR merged for 1.13.0 is causing the DB instance to recreate due to the way the attribute is defined (even if the end result value isn't changing).

See that PR here #236 (review)

https://sweetops.slack.com/archives/CB6GHNLG0/p1733335202382489

Expected Behavior

It should not force a recreate of the resource. If it forces a recreate, it should be a MAJOR bump.

Steps to Reproduce

Set version to 1.13.0, will force replacement

Screenshots

No response

Environment

No response

Additional Context

No response

[GabisCampana]

@aknysh

[chrisforrette]

Hi there!

I just wanted to share that this is still an issue for me. I've been trying to upgrade from version ~1.3 and running into this with the latest 1.x and 2.x. Running terraform plan show it wants to recreate the DB cluster instances citing the identifier as the reason even though it's unchanged (~ identifier = "unchanged-instance-name-2" -> (known after apply) # forces replacement). I went version by version and ran plans and found that this was introduced in v1.10.0, which makes sense as it looks like that version made a shift to setting the identifier from a dynamic source.

I attempted several workarounds with no success—passing in cluster_identifier, running terraform state rm on the aws_rds_cluster_instance resources, etc.—but the plan kept wanting to replace them. For now I'm just sticking with 1.9.1. Are there any known workarounds for this?

Happy to share any other information if it's helpful. Thanks!