Details
-
Bug
-
Status: Closed (View Workflow)
-
Minor
-
Resolution: Fixed
-
all
-
None
Description
When using a PrepareStatement executing on a Aurora replica, and failover append, master can be use temporary.
When a replica is up again, the prepareStatement must fall back again using the replica connection.
Depending on failover parameters, this may take a long time (seen on test about 13~17 minutes with the folowing parameters : validConnectionTimeout=120
socketTimeout=10000
failoverLoopRetries=120
connectTimeout=250
loadBalanceBlacklistTimeout=50)