Hyper-V Replica allows you to do:
- Test failover: bring online copies of the replica VMs in the DR site, ideally on isolated virtual switches to test without interfering with production systems.
- Unplanned failover: An unexpected disaster such as an earthquake, tsunami, or fire, wipes out the building/site. You failover the VMs, losing maybe up to 10 minutes of data or more (replication every 5 minutes, on a best effort basis depending on the WAN)
- Planned failover: VMs are shutdown in the production site, the HRL is flushed and replicated to the VMs in the production site. VMs are started up in the production site.
Folks in the USA south are fully aware that a hurricane is bearing down on them. This is not a new experience for New Orleans. Mandatory evacuations are in place in some parts. This is exactly the scenario when a planned failover is best used.
Weather forecasts have predicted the onset of the hurricane and a likely trajectory. Now is the time to perform that planned failover and have the VMs up and running in the DR site. Benefits?
- It is a smooth transition with no data loss
- The systems will be up and running in the DR site before the chaos of the evacuation and disaster
The dependency on the virtualisation administrator is greatly lessened with a planned failover already done. To be quite heartless about it, the impact on the company will be lessened if the virtualisation admins are killed or are trapped by the disaster.
It is possible to orchestrate/order the VMs. Some want to have a go about the lack of System Center 2012 SP1. Pfft! System Center 2012 is nice, but don’t need it. A few lines of PowerShell and I have all the orchestration that I need. I’ve already written/proven it and I’m no PowerShell guru.