0

We are looking at different options to allow us to stay up and running when our primary datacenter (DC1) goes down for any reason. What is the main difference in having a offsite server updated all the time using log shipping as a warm standby? (meaning we would manually redirect the DNS if DC 1 goes down or is offline to DC2).

Versus running a Hyper V cluster and setting it up to automatically fail over to the DC 2 when DC 1 is down? Although its not clear how the traffic gets redirected in real time unless we were running some sort of load balancer in the cloud?

I assume the Hyper V cluster option will provide us minimal downtime if it is automatic? Then do we even bother with sql server log shipping? It seems that these two things overlap in certain ways and we don't need to do both?

I am aware that the Hyper V cluster requires shared discs of some sort (was looking at iscsi target as a san option for server images in this case), but sql server log shipping also seems to require a witness server, so either way we have to introduce more hardware to manage.

DB is 30Gb running on SQL 2008 R2 , web is 2008 R2 (both planned to upgrade to 2012 in next few months).

1
  • You need to go back to reading documentation. Hyper-V cluster is not a solution (VM's still go down - that is not HA), and depending how you set it up does not require anything shared at all. How you set this up is VERY depending on what you do (application) and we dont have enough info for that. If that is a simple "sql server + web" then also look at what SQL Server has. Note MIRRORING, not log file shipping.
    – TomTom
    Jan 6, 2013 at 18:06

1 Answer 1

1

To answer your question, you need to figure out exactly what you were looking for. Let me define some terms.

High Availability (HA) to me (though not to TomTom, commenting on your question) means that no single failure can take your VM permanently off line. In the context of Hyper-V, this is a host cluster where the failure of one physical node will result in the VM being restarted after the crash on another node.

Fault Tolerance (FT) to me means that a virtualization-layer failure (in software or hardware) will not be observed in the VM. The VM will continue to run. VMware and Xen and presumably other systems can do this, with many caveats, usually most notably a very low limit on the number of virtual processors in the VM. Hyper-V does not provide FT, though you may find a server hardware vendor who can supply a system that will solve the hardware-layer problems.

Disaster Recovery (DR) to me means that the failure of an entire datacenter (or large subsection thereof) will not keep your VM offline. Hyper-V in Server 2012 provides "Hyper-V Replica" to address this scenario.

If you want automatic failover, then you want a "stretch cluster" of Hyper-V hosts with HA VMs where some nodes of the cluster are in your primary datacenter and some are in your secondary datacenter. This will require shared block storage (usually a SAN) with a presence in both datacenters.

Often, though, you don't want automatic failover, either because you want to be very certain that your primary datacenter is really off line before moving your services to another datacenter or because you need to do some level of reconfiguration when the VMs move. Hyper-V Replica is targeted at this situation. You can make it more automatic by scripting the failover, if you like, but that's something you would overlay on top of it.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .