Details
-
Task
-
Status: Closed (View Workflow)
-
Minor
-
Resolution: Not a Bug
-
None
-
None
Description
When using AWS Aurora with a read replica and the load balancing provided by the driver, you're faced with potential inconsistency because of the replica lag (an insert is not instantly reflected on the read replicas). This means that you can insert an object, immediately retrieve it on a readOnly=true connection and get null.
I understand that in some cases this lag is acceptable and the load balancing advantages outweigh the consistency. However, there should be an option so that all queries, regarless of the status of the readOnly flag, are routed to the master instance while retaining the failover capabilities of the driver.
Attachments
Issue Links
- relates to
-
CONJ-751 Support Aurora mode with a Master only connection
- Closed