First of all, Thank you for a great job on creating OpenDistro for Elasticsearch. It has benefited a huge number of organizations.
I would like to receive a feedback from the experts on the Cross Cluster replication. My current deployment architecture for ES installation is a multi-node cluster on kubernetes running 3 master, 2 Data and 2 Client nodes and storing the data in a persistent volume (glusterfs) at a local cluster.
For DR/HA cluster, I would like to understand how community users are implementing the Cross-Cluster replication.
Here are the options that I am considering:
Running both sites to store data in Persistent Volume (realtime): With this approach, I will basically be getting data from kafka Queues where I can fetch data using unique group/consumer at each site.(Ref)
Network File System(NFS): Using an external network file system which will be accessed with Pods(in K8s) and be available to both the sites.
Snapshot & restore using ES Curator
Can some one please enlighten me if you think above options have proved well or have any other recommendations?