Some more information:
I just did some experiments and it is taking up to 15 seconds to fail-over
on a primary step down. This seems ridiculous. Does anyone know if this is
configurable?
On Monday, October 8, 2012 1:01:09 AM UTC-6, Cwolf wrote:
> Hello - It seems that the replication heartbeat is not configurable. In
> 1.8 it was (did not check 2.0). Are there any details on how heartbeat and
> fail-over timings work? What is the timing on fail-overs now (it seems
> fast - but need to understand). We are looking to get sub second fail-over
> working in an environment where all machines are on a single switch.
> Thanks