Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

We currently run a Hyper-V server on a non-domain W2008 server, and are looking to build a failover solution. Because all the (6-10) VM guests are linuxes, and our datacenter space is 90% non windows, we have never felt the need for AD domains etc.

My initial plan is a second server (either W2008R2 or W2012 standard), running Hyper-V server.

There is no shared storage / SAN in place.

The VM's are pretty static - portals and web servers, with database located elsewhere, off the VMs, and so we could reasonably easily cope with week-old backups copies regularly from master VM to secondary VM.

Option 1 is the above setup - if VM Server goes bang, spin up the copies on secondary.

Option 2 is implement Failover Cluster, so if VM Server goes down, VM Secondary will just take over

I would really appreciate some advice on the way forwards with this - the more I read the more confused I get.

share|improve this question
Can I just ask, if you're a non-MS shop, why Hyper-V at all? Would one of the other big names not suit you better? – Dan Jan 8 '13 at 10:35
Are you running the free Hyper-V server? – pauska Jan 8 '13 at 10:40
Yeah. Currently server running VM's is W2008 Std (not R2), so an upgrade looks likely to be needed anyway. Half an hour's downtime while we spin up the guests on the standby is not a crisis. But automagic transfer from one to the other would be nice (especially if it's 3am and I'm tucked up in bed). But I feel very vulnerable with one VM server - if the power supply goes bang, or the processor explodes, I am stuffed. – David Shields Jan 8 '13 at 10:47
The question on why use Hyper-V is still a valid question. If you're not a Windows server shop in general, why go Hyper-V? – Rex Jan 10 '13 at 6:37
up vote 0 down vote accepted

No, but why do you NEED that?

I mean, can you not handle the high uptime requirement on the VM level? Have 2 sets of vm's (one on each server) and then do the HA part in the software.

  • Databases should support that out of the box of with some third party. SQL Server for example can do that (mirror, high availablility)
  • Web machines should support that too.

Basically, assume a Hyper-V server CAN crash, and handle the failure WITHIN the VM layer, similar like you would do it if that all would not be virtual but every VM would be a physical machine.

Any HA for Hyper-V requires a domain as this is what windows clustering uses. And that still leaves issues (corrupt images). But that is not HA anyway (requires restart) and there still is the app layer level.

share|improve this answer
This doesn't really answer the question at all. Using virtualisation for redundancy at the hypervisor level is completely reasonable and is half the reason that companies go the virtual route. – Dan Jan 8 '13 at 10:34
Ah, yes, but then this has a "get a domain in place" requirement. Plus this is not a real solution in many scenarios - I personally go down BOTH ways. Some machines are like that, some are internally handled, mostly that which are important. But anyhow, Hyper-V HA requirements ARE laid out and they DO require shared storage and a domain. Point. Fight requirements - your problem. – TomTom Jan 8 '13 at 10:35
@TomTom Hyper-V in WS2012 does not require shared storage for failover clustering. – pauska Jan 8 '13 at 10:45
@pauska Please show references, because this is BULL. Either shared storage, or replication which has a time delay and not HA but business continuity. Hyper-V does not support real time replication. I unerstand you do not know Hyper-V. – TomTom Jan 8 '13 at 10:47
Not entirely sure I understand what you are saying here (certainly due to my lack of knowledge on windows systems - last time I was a sysadmin for domains was NT4!). Optopn 1 would have clones on each guest on each vm server, so standby guests could not spin up as their (static) ip's could clash. – David Shields Jan 8 '13 at 10:50

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.