Description
Currently there is a validation step for connector names which prevents sink connector consumer groups from colliding with the worker group.id.
There is currently no such validation for consumer.override.group.id that would prevent a conflicting connector from being configured, and so it is possible to misconfigure a connector in a way that may be damaging to the workers themselves.
Reproduction steps:
1. Configure a connect distributed cluster with a certain group.id in the worker config.
2. Configure a sink connector with consumer.override.group.id having the same value as in the worker config
Expected behavior:
1. An error is returned indicating that the consumer.override.group.id is invalid
2. The connector is not created or started
Actual behavior:
1. No error is returned, and the configuration is otherwise valid.
2. The connector is created and starts running.
Attachments
Issue Links
- is related to
-
KAFKA-3829 Warn that kafka-connect group.id must not conflict with connector names
- Resolved