vaso - Fotolia
Speed Hyper-V replication process by seeding replica VM
Hyper-V replication lets IT create secondary, offline virtual machines that can be used to recover operations following an outage of the primary VM.
Hyper-V allows administrators to protect virtual machines (VMs) by maintaining offline VM replicas in the data center or in a remote facility. If something were to happen to the primary copy of a VM or to the data center as a whole, the replica VM could be quickly brought online to service production workloads.
Although Microsoft makes it relatively easy to create a VM replica, there are a number of decisions an administrator needs to make. Some of the most important ones pertain to the initial replication process. So what is the best way to copy the data from a production VM to its remote Hyper-V replica?
The easiest way to perform the initial Hyper-V replication process is to replicate the VMs across the network. While it's easy to do this, it isn't always the best option, especially if:
- You have a lot of VMs that need to be replicated. For example, if a VM is replicated across a WAN link, initial replication could take weeks to complete depending on the size of the virtual machine.
- You need to replicate one or more very large virtual machines. The initial replication process has a tendency to fail if the VM contains more than a terabyte of data. When a replication failure occurs, the replication process can be restarted from the point of failure and will eventually succeed, but babysitting the replication process so that it can be restarted each time it fails is a waste of an administrator's time.
- You have slow or unreliable connectivity between the primary and secondary server.
Seeding the replica VM
But other options exist. Rather than performing the initial Hyper-V replication process across the wire, Microsoft gives administrators the option to seed the remote Hyper-V replica. Seeding involves writing data directly to the replica server so that only the deltas have to be transmitted across the wire.
There are two ways to seed a replica VM:
- Use an existing virtual machine on the replica server as the initial copy. To use this seeding method, you need to back up the VM you want to replicate and then restore that backup to the replica server. Hyper-V then allows you to treat the newly restored backup as the initial replica. Since the virtual machine's operating system and the bulk of its data now reside on the remote host, an administrator need only designate the VM as a replica and Hyper-V will begin synchronizing the primary VM to the replica VM.
- Microsoft also has a method that works similar to the backup/restore method, but does not require backup software to be used. This method allows you to export the virtual machine to removable media and then import the VM directly into the replica host.
You can see the available initial replication methods in Figure 1.
As you look at Figure 1, you will notice an option to schedule the initial Hyper-V replication. Scheduling can be used if you are creating the replica from an existing VM or sending the initial copy over the network, but you can't use the scheduling option if you are importing the VM from removable media. Scheduling allows you to initiate the initial replication during off-peak hours so you can avoid hardware resource consumption during peak business hours. It is important to remember that Hyper-V only allows you to schedule the time when the initial replication begins. There is no mechanism to pause the initial replication process if it runs long.
Each of the processes outlined above has its advantages and disadvantages, but there isn't one method that is the best choice in every situation. You will therefore need to consider which one will work best in your own organization.
How to use Azure as a replication target for Hyper-V
How to provision storage for the replication feature in Hyper-V
Situations best suited to host applications in Hyper-V Replica