Skip to content

v1.2.22 release of dynamodb-controller deleted preconfigured tableReplica #2400

@mattzech

Description

@mattzech

Describe the bug
We had a team who created a DynamoDB Table in production last year which was then configured with a replica in us-east-1 via AWS (since this was not supported in Table spec yet). When the new dynamodb controller release went out to this cluster, the generated Table resource had its spec.tableReplicas set to null, which then deleted the pre-created replica in us-east-1, causing downtime for this application
Steps to reproduce

  • While using v1.2.22, adopt a DynamoDB table with a table replica already configured in another region
  • Upgrade to v1.2.22 and check AWS for replica table (observed deletion in cloudtrail)

Expected outcome
Somehow there needs to be better protection for Tables where this field is not set

Environment

  • Kubernetes version v1.29.11
  • Using EKS (yes/no), if so version? no
  • AWS service targeted (S3, RDS, etc.) DynamoDB

Metadata

Metadata

Assignees

No one assigned

    Labels

    service/dynamodbIndicates issues or PRs that are related to dynamodb-controller.

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions