Drift Detection issue for Classic property under AWS::Redshift::Cluster resource

[amedeshm]

Name of the resource

AWS::Redshift::Cluster

Resource Name

No response

Issue Description

The Classic property for the AWS::Redshift::Cluster resource is marked as drifted when its value is specified as false.

Expected Behavior

The Classic property should be IN_SYNC or NOT_CHECKED status.

Observed Behavior

The Classic property is marked as REMOVE with expected = false actual = -

Test Cases

Steps to reproduce -

1. Create Stack using following template -

Resources:
    myCluster:
      Type: 'AWS::Redshift::Cluster'
      Properties:
        DBName: mydb
        MasterUsername: master
        MasterUserPassword: xxxxxxxxxx
        NodeType: ds2.xlarge
        ClusterType: single-node
        Classic: false

2. Run drift detection
3. View drift results

Other Details

No response

[shwetayakkali]

classic property indicates A boolean value indicating whether the resize operation is using the classic resize

sometimes customer may remove this value when modification/update in terms of resize is done, this isn't a core param for cluster, added to ease cloudformation template to indicate classic resize. describe-clusters used in drift will not indicate this property and hence this property once used in updating should be removed. We can resolve this ticket imo. Will get in touch with docs team to document this case for Redshift cluster drift.

[shwetayakkali]

Filed a ticket downstream to ignore this property.

[rgoltz]

Hi @shwetayakkali - Thanks a lot for the provided information. Unfortunately, we still see a drift reported if we set Classic property to false in CFN template. Having this said, the property seems not to be ignored (tested in eu-central-1).

Could you please re-open this issue here? (or should I create a new issue with reference to this?)

[greg5123334]

Confirmed. This behavior is consistently reproducible. The 'classic' property is reporting false-positive drift, regardless of the value.

Test code:

    key = kms.Key(self, "MyKey",
                  removal_policy=RemovalPolicy.DESTROY,
                  )

    cluster = CfnCluster(self, 'Cluster',
                         cluster_type='single-node',
                         db_name='dev',
                         master_username='bevelvoerder',
                         master_user_password='Wagw00rdEen',
                         node_type='dc2.large',
                         encrypted=True,
                         kms_key_id=key.key_id,
                         classic=True,
                         )
    cluster.apply_removal_policy(RemovalPolicy.DESTROY)

Recommend reopening this ticket for further consideration.

[rgoltz]

@prerna-p Cloud you please re-open this issue, please?

[rgoltz]

@CloudChoom Lucas / @kanitkah Harshu - Could you please re-open this issue, since this bug is still in place? The wrong status of this github issue affecting our tracking of false-positives of drift detection. Thanks!

[kanitkah]

@rgoltz We will get this ticket reopened.

[shwetayakkali]

Classic property is a writeOnly property, moving it under writeOnly will not have drift.

please confirm.

[rgoltz]

@shwetayakkali I assume your question is more related to the underlying CFN implementation, which may be addressed in a linked issue under https://github.com/aws-cloudformation/aws-cloudformation-resource-providers-redshift ? (this worked really good with the RDS team in the past)

[mashuaihh]

Verified that it's no longer an issue