EMR recreates resources when master_instance_group_instance_count is greater than 1

[homiakos]

Describe the Bug

If you specify a value greater than one, then the cluster is recreated.

I think the problem is in the terraform itself:
hashicorp/terraform-provider-aws#10446

  But now it would be nice to warn users about this
The problem is reproduced with
core_instance_group_ebs_volumes_per_instance
master_instance_group_ebs_volumes_per_instance

Expected Behavior

To warn that now it is worth using only the default value, otherwise the cluster will be recreated

Steps to Reproduce

Steps to reproduce the behavior:

1. set config
   core_instance_group_ebs_size = 32
   core_instance_group_ebs_volumes_per_instance = 2
   master_instance_group_ebs_size = 32
   master_instance_group_ebs_volumes_per_instance = 2
   2 . Run apply twice

2. Second apply see:

       - ebs_config { # forces replacement
           - iops                 = 0 -> null
           - size                 = 32 -> null
           - type                 = "gp2" -> null
           - volumes_per_instance = 1 -> null
         }
       + ebs_config { # forces replacement
           + iops                 = 0
           + size                 = 32
           + type                 = "gp2"
           + volumes_per_instance = 2
   

[3h4x]

Can you describe what was configration first and then to what have you changed it?
I can only guess that the problem is that you changed volumes_per_instance from 1 to 2 which is triggering recreation of cluster.

EDIT: I've read the issue related - bug in terraform provider.
Thanks for creating that issue! It will increase visibility over EMR issue.