Ne04J Causal Cluster architecture in AWS


(Rxxk) #1

Hello, We are exploring Ne04j for one of our requirements. Have few architecture level questions on Ne0j4 deployments in AWS in terms of High Availability and Disaster recovery.

  1. Is it possible to deploy one core server in this mode and use AZA and AZB for HA and Region2AZA as DR ?

Region1- AZA – Coreserver1
Region1- AZB – Coreserver2
Region2- AZA - Coreserver3

  1. As writes get acknowledged with majority members in the cluster( In this case 2), will Ne04j pick the members with less latency automatically ?. In this scenario, will writes get ACK, as soon AZA and AZB acknowledges back?
  2. Is the above behavior is by default or do we have to configure something at the cluster level?
  3. Can we convert a read replica into a core member?, if yes, please point me to some documentation.
  4. What other patterns are feasible to deploy Cores in AWS Multi-AZ for HA and DR in out of the region with minimal core servers?

Appreciate if you can shed some light on these questions.

Regards
Rx