Does Confluent provide a CDC connector for Oracle ? Does it need debezium ? We plan to use Kafka Worker clusters. Is there buffering involved in the worker ? Is it correct to assume that once the worker pulls the records it will be delivered ‘Exactly-once’ to the broker ?
This is a local Kafka setup.
I came across Sign in to your account
Can I look at a diagram showing Oracle LogMiner and this connector ? Does it use LogMiner ? Does it buffer the data inside the connect Cluster when it cannot deliver to Kafka ?
No, Confluent doesn’t have a primary developed CDC connector; Debezium is a RedHat project
Yes, it uses logminer or Xstreams
If data cannot be produced to Kafka, then records will be dropped by the producer api, or the connector will stop with an exception, not cache data indefinitely
Thanks. I came across a connector developed by A2 systems. That is in Confluent website. I think. So any Kafka Connector cluster whether it is connects to MQ or a database have to be monitored properly. I believe that is what you mean ?
Is there any flow diagram showing the failure modes ?
The title of this post says Debezium, though… In any case, Confluent Hub is just a catalog. There is a range of support models for connector projects from not at all (community developed, but not maintained), to enterprise level support.
I’m not familiar with this A2 connector or how gracefully it handles exceptions, but yes, monitoring is recommended and that’s what JMX or the /status endpoint of the Connect API are for
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.