Details
-
Improvement
-
Resolution: Fixed
-
Major
-
None
-
None
-
1
Description
The Kafka connector currently uses the Java SDK's default retry strategy, which considers some kinds of failures to be unretryable (particularly the kinds of failures that occur during a Couchbase failover/rebalance).
Instead, we want the connector to retry writes that fail for any reason until a user-configured timeout duration expires.
Attachments
Issue Links
- is duplicated by
-
KAFKAC-217 Sink connector fails if any of the couhbase node is unreachable for some time
- Resolved
-
KAFKAC-140 Test and document behavior when Sink write fails
- Resolved