AMQ journal recommendations with remote disks

Investigate and propose changes on Docs, Tests and broker configurations to improve the customer experience when the broker is using remote disks eg NFS, SAN ,...

Our current reccomandation for *nix users using remote disks, regardless any performance target, is to still use ASYNCIO journal, but this requires customers to take care of providing a reliable and stable remote disk behaviour.
A remote disk misconfiguration can make ASYNCIO I/O operations to pile-up on kernel for undefined amounts of time, depending on the severity of a network glitch/outage, making the broker to time-out (with an I/O critical error) and shutdown itself.

The motivation of this issue is to improve the user experience with the broker, proposing/documenting and testing a solution that will met the required reliability while preserving enough performance to meet the target of most customers.