We want to use Hyper-V Replica to copy VMs to a DR site (connected via 10GB fiber). I have found that VM write speeds to VHDXs (small and large) are drastically reduced. I'm attaching an image of benchmark testing. This is the behavior I'm seeing in the VM, not the host.
This behavior has been seen on all VMs with a variety replication partners.
Here's an example of the configuration.
VM: Windows Server 2016 Standard | Gen 2 | 8->32 GB RAM | 12 CPU | SCSI Controller
Hyper-V Environments:
- Windows Server 2019 Standard | 382 GB RAM | DA two tier storage (SSD->HDD)
- Windows Server 2016 Datacenter Cluster | 383 GB RAM | CSV on Equallogic via 10GB private network
- Windows Server 2019 Datacenter Cluster | 447 GB RAM | CSV on StorageCenter via 40GB private network
Is Replica the appropriate tool for disaster recovery? If not for the performance, it seems that it should be ideal.