← Back to team overview

nova team mailing list archive

RPC and fault tolerant configurations for RabbitMQ


Currently Nova RPC deals only with intermittent network connectivity and I wonder whether configurations like RabbitMQ clustering and active/passive replications have been looked at to provide fault-tolerance.

I have seen a new blueprint on "Refactoring configuration for message queue" submitted by Jay Pipes, so I am asking...is it appropriate to write a new blueprint explicitly for fault tolerance? Should that be done within the same blueprint or implemented outside OpenStack altogether?

Many thanks,

Follow ups