Clustering and/or Hyper-V replica?

We currently have one big Dell server T610 physically hosting 4 Hyper-V VMs all Win2008R2. Running smoothly for years. We now need to purchase a new Dell server since Dell won't renew the existing service contract's 'end-of-life' T610.

We're a small biz at a single site. We've always been nervous about the ONE physical server. I think we need two servers one-site for full redundancy. If one box dies one day, we need to be back up in short time.

We need advice on what we should be looking at. Windows Server Clustering? Hyper-V replia? We simply need protection from our current single point of failure exposure.

Solution 1:

I would recommend you to use your old hardware in another way. I would purchase a new Dell server with the similar specs and create a Highly Available cluster. You can take a look at StarWind Virtual SAN that, as I know, can provide you with Highly Available storage. You will be able to create a redundant cluster with two nodes. As another option, you could take a look at what HPE offers.

Solution 2:

Few words about Hyper-V Replica as it's one of the most frequently deployed scenarios.

You can PowerShell automate Hyper-V Replica no problem!

https://blogs.technet.microsoft.com/keithmayer/2012/10/05/automated-disaster-recovery-testing-and-failover-with-hyper-v-replica-and-powershell-3-0-for-free/

There are some caveats however:

1) There's no protection against a brain split, so you might want to run your orchestration site somewhere.

https://docs.microsoft.com/en-us/azure/site-recovery/site-recovery-faq

2) Hyper-V replica in a pre-WS2016 brings in a 2x write amplification gap. WS2016 is OK and third-party software using own changed block tracker, think of Veeam and StorageCraft don't have this issue.

https://forums.veeam.com/microsoft-hyper-v-f25/hyper-v-2012-r2-replica-vs-veeam-off-host-replication-t16374.html

3) There's some data loss on Hyper-V Replica failover and some downtime. You may be OK with them and maybe not, it's up to you completely!

https://redmondmag.com/articles/2014/12/01/hyperv-replica-for-disaster-recovery.aspx

Think about your RTO and RPP #s and your budget before you move on with your implementation.

Solution 3:

A Hyper-V Failover Cluster will afford you automatic failover and failback in the event of a host failure. This requires shared storage between the cluster nodes.

Hyper-V Replica affords you manual failover and failback in the event of a host failure and doesn't require shared storage.

It's really a matter of what your budget will allow and what type of failover you require.