Bug report

What Debezium connector do you use and what version?

Debezium Oracle connector version 3.0.7.Final

What is the connector configuration?

oracle-connector-config.txt

What is the captured database version and mode of deployment?

Oracle Database 12c Standard Edition Release 12.1.0.2.0, hosted on-premise.

What behavior do you see?

In some cases, Debezium logs the start of a new mining session ("Starting mining session") and remains stuck at this waiting for a response from Logminer forever, no longer capturing new change events or updating its metrics. No exception is raised and both the connector and task remain in RUNNING state even though it is not functioning normally. The property "database.query.timeout.ms" is set to 10 minutes by default, however debezium never timeouts. I have attached logs bellow.

What behavior do you expect?

Debezium should eventually timeout after waiting for too long for a response from Logminer, raising a clear error (and thus moving to Failed state). It would make it easier for external monitoring processes to detect that debezium has indeed run into an issue and send alerts/recover.

Do you have the connector logs, ideally from start till finish?

all-pods.log

How to reproduce the issue using our tutorial deployment?

The issue happens rarely and I don't know how to reproduce.