How to manage ksql's queries when failover with cluster linking?

Stream Processing ksqlDB
1

Hi team,

I tried to migrate ksql’s queries from DC1 to DC2 with CP that set up cluster linking. So I need some help with the design solution for failover ksql’s queries.

ksql’s queries : stream transaction join with table of master data
CP details: Set up as Active-Passive and normally DC1 : active / DC2 : passive
my question:
When promoting DC2 as new active we already sync topic, schema and consumer offset with cluster linking from DC1->DC2.
But in part of ksql’s queries, How do mirror all queries and all states from DC1 to DC2 to continue process data exactly where they left off without missing or duplicates?

thanks
May.r

2

The DR strategy for KSQL using cluster linking is documented in Cluster Linking for Failover and Disaster Recovery on Confluent Cloud | Confluent Documentation

TLDR: Recommendation is to have a standby replica running in the DR region and mirror only the input topics. You can also mirror the internal topics but rebuilding the state will be time consuming