Erick, even if we give service as a contact point datastax driver will calculate IPs of topology to connect the cluster. So I think as John proposed in answer to similar question, after driver initialization in case of total outage of Cassandra cluster, bouncing of application is the only option.
Regards Manish On Tue, Aug 11, 2020 at 4:42 AM Erick Ramirez <erick.rami...@datastax.com> wrote: > Use the service as contact points, not the IPs since they're ephemeral. > Even when all pods get replaced, they will still be accessible via the > service. Cheers! > >>
