Hyper-V replica was introduced in Windows Server 2012 as a new feature for Hyper-V environments. It has great potential and is very helpful in virtualized systems. If you are looking for an introduction to Hyper-V replica and want to learn more details, keep reading. I’m going to present the 5 most important things to know about Hyper-V replica.
Purpose of Hyper-V Replica
The purpose of replica is to provide the best RTO possible. RTO (Recovery Time Objective) is a value, which states how much downtime you can allow your system without having a significant impact on business. It sets a time when services should be back up and running.
Obviously, in case you restore from a backup, additional time for extracting one VM’s data from a compressed and deduplicated backup file and transferring that data back to the system must be accounted for.
With Hyper-V replica, however, the recovery process is done in a matter of seconds. Due to replica design, replica files are waiting in a ready-to-be-used state on a DR site and can be used when it’s needed.
Versions
Unfortunately for some administrators, the replica feature wasn’t released until Windows Server 2012 and then improved in Windows Server 2012 R2. It’s not available on Windows Server 2008 or Windows Server 2008 R2.
Besides that, a replication server (a server that hosts replicas) should have the same or newer version as the production server. So, it’s not possible to re-use a 2012 server when performing replication from 2012 R2. The old one would have to be updated as well.
Enable Hyper-V replica
The replica enablement process is just easy and can be done within a few minutes. You should select “replica” in the VM settings, specify the “DR” server and connection parameters, select any of the VM disks and set the retention policy and initial replication method. For the last operation, Microsoft asks you to select between instant replication over network, copying data to external drive (and moving it to DR side manually) and seeding from an existing VM on DR site. As soon as initial replication is done, changes in the primary virtual machines are transmitted over the network to the replica virtual machines periodically.
Retention Policy and scheduling
Replica mechanisms are not meant to be completely “live”. There are some delays according to replication settings. In Windows Server 2012, the interval between two replica iterations is set to 5 min and can’t be changed. In Windows Server 2012 R2, you can select between 30 sec, 5 min and 15 min. The scheduling isn’t very flexible. If there is real need not to interrupt system within production hours, replication shouldn’t be used due to this low time interval since it could impact the system.
Regarding to retention period, in a 2012 server, the maximum number of restore points is limited by 15 hours. This means that you would go back to 180 points from the last 15 hours and switch to any of them when it is needed. In Windows Server 2012 R2, the number is increased up to the last 24 hours which gives roughly 2,880 points (24 hours * 60 min/hour * 2 points per minute)
Planned failover or unplanned
It’s possible to make a planned failover. Find a maintenance window, turn off VM(s) and activate “planned failover” mode. Hyper-V will turn on associated replicas, make them work instead of the original VMs and switch replication settings to opposite. New information from working replicas will be transferred to source VMs, temporarily making them replicas.
Use of unplanned failover comes in place when you have a disaster with the system and want to migrate to an unaffected replica server.
Bonus
The functionality of Hyper-V replica might be extended with Veeam Backup & Replication. Using Veeam makes retention strategies flexible and simple, providing more options to configure it. Besides that, Veeam allows to deploy a dedicated off-host proxy server in the source site, and all the necessary data transformations (such as compression and block filtering) are performed on it which helps reduce unwanted overhead on the source Hyper-V host during replication. Regarding transferring data process, Veeam considers the capacity of network connection and suggests enabling traffic throttling, TCP/IP multi-threading or optimizations for high-latency links. Furthermore, Veeam offers a bunch of recovery scenarios like file/item/VM file level restores.
And last but not the least, replication features are already included in this Veeam backup solution, so it helps to reduce the costs of protective systems in the end.
In conclusion, Hyper-V replica is a great feature that reveals yet another area of virtualization and makes it even more attractive to administrators. Make a setup, play with it and keep in mind that replica is not intended to be a replacement for backup solutions, but always complements them.
