Can anyone help?
Thanks,
Chen
Hi Chen,
What is the consistency level of your requests?
I'm pretty sure the cpp driver is defaulted to LOCAL_ONE, so if your client is in the "failed" datacenter that's a problem. Also because you have a total of 6 replicas, a simple QUORUM consistency would fail because a QUORUM in your case is 4, but you can only get a maximum of 3 nodes to respond.
Hi Chen,
I have posted the same question but didn't get any answer. Later I got the answer from datastax enterprise support that cpp driver doesn't have a failover retry policy. We need to do an application retry for these scenarios.