Open soudaburger opened 1 year ago
This would also help in the situation of: #481
There is some discussion about this being done, however this requires additional tooling on top of what we have and will probably have in this chart. Please stay tuned.
What would you like to be added?
I basically would like to be able to deploy a second set of kafka pods in a separate cluster in a separate region to allow them to replicate data between my regions. This would ensure that even in the event of a failover DR event, I don't actually have to modify any deployments of services using kafka, as both regions effectively become active.
Example:
redpanda
helm chart with stateful set tous-east1
as a starting cluster. Let's assume 3 nodes.redpanda
helm chart with additional statefulset tous-west1
as an "expansion". Let's also assume 3 nodes.This would provide me with 6 Kafka nodes, all able to take traffic and replicate accordingly. I could, in theory, allow my services to write to either region and support running a hot/hot architecture, which is desired.
I assume that I would need to have the following set up in order to make this functional:
Why is this needed?
This would allow me to run a hot/hot cross-regional deployment on separate clusters to avoid certain catastrophic failure scenarios.
JIRA Link: K8S-37