What is the correct way we should power down a 2-node hyperconverged setup when the power fails and we are on UPS power? The APC software has options to migrate VMs, and also to shutdown VMs, and also to shutdown the physical hosts. The main concern is with the write buffer getting de-staged in time, so how can I make that happen as soon as possible?
Should I just do the typical sequence of 1) shutdown VMs, 2) shutdown hosts? And then StarWind will do the right thing to de-stage the write buffer?
Or do I need to do something like 1) at power outage, immediately migrate VMs from Host1 to Host2, 2) shutdown Host1 ASAP (this should disable write buffer on Host2, I think), 3) sometime later shutdown VMs on Host2, 4) shutdown Host2.
What is the best practice here?
The Latest Gartner® Magic Quadrant™Hyperconverged Infrastructure Software
