If you start multiple Kafka connector instances on the same topics or partitions, what can occur?

• A. Multiple copies of the same row might be inserted
• B. Data would be missing
• C. Connector will coordinate and prevent duplicates
• D. The topics will be ignored

Answer: (A)

When multiple Kafka Connect instances run on the same topics or partitions, the same records can be processed by more than one instance. If there’s no strong cross-instance coordination or exactly-once delivery, those records may be written to the sink more than once, so you can end up with duplicates in the target system. In other words, you might insert multiple copies of the same row. The other outcomes—data going missing, automatic coordination preventing duplicates, or topics being ignored—don’t reflect the typical risk when several connectors operate on the same data without a global coordination mechanism. To avoid duplicates, you’d need to use a single coordinated connect cluster, enable exactly-once semantics where possible, and ensure the sink can handle idempotent writes or upserts.