Hi All, I am planning to use phoenix in production and one of the basic requirement is to have a DR (disaster recovery) stack for the Hadoop cluster, which will reside in a different data center.
In Phoenix I am using salted tables with mutable global indices and GZ compression, I was analyzing if I can use Hbase (Master-slave) replication to replicate data stored using phoenix. So far I think it is completely doable if I replicate main table, all associated index tables, along with the phoenix System.catalog , system.sequence, and system.stats table. Only place I think this replication can make data inconsistent is if any region split happens on a Slave cluster. Am I right in thinking that or have I missed any aspect of it ? If any of you has a similar case study please share it. Thanks, Dhaval
