We wanted to take a minute and respond to the recent blog post by the Microsoft Exchange team (Answering Exchange Virtualization Questions and Addressing Misleading VMware Guidance). In the post, Microsoft expresses concerns around particular VMware guidance for deploying Exchange Database Availability Groups (DAG) in conjunction with VMware HA. We’d like to provide our perspective on this configuration and explain why it isn’t reckless (as Microsoft suggests), but all to the contrary can help our customers get greater value out of virtualizing Exchange.
The proposed solution consists of deploying Exchange DAG on a VMware vSphere cluster, and complementing DAG with VMware HA to maximize availability.
To clarify some of the apparent confusion on VMware HA in the Microsoft blog, VMware HA has nothing to do with VMotion or DRS. VMware HA is a very simple mechanism that automatically reboots a failed virtual machine on any available host in the vSphere cluster in the case of a host failure. From an application standpoint, the behavior of VMware HA is equivalent to a simple power-on of the machine on which the application is running. It is completely transparent to the OS and the application.
In the physical world, when a DAG node fails, an admin will eventually power-on that node to restore the original availability protection. VMware HA does exactly the same thing, but automatically. Seems to us like that’s a pretty reasonable thing to do, and many of our customers agree and are happily using this in production. Why? For example, in a deployment with 2 database copies, if one of the DAG nodes were to go down due to a host failure, Exchange would be left ‘unprotected’ until the failed DAG node could be brought back up. That could take quite a bit of time, relying on an administrator to detect the failure, and manually power on the failed node. VMware HA automatically takes care of that for you in the shortest time possible to restore the DAG availability protection.
Does this solution increase the cost of deploying Exchange? No! Compute requirements are unchanged, since VMware HA doesn’t rely on failover instances. VMware HA also does not impact overall storage requirements.
Does this solution increase complexity? Our customers think our solution makes Exchange a whole lot simpler. When a DAG node fails, you no longer need to manually detect the failure and bring the node back up – VMware HA does that automatically. And VMware HA doesn’t require any OS- or app-level configuration changes.
As to the support situation, we acknowledge that this is not an officially supported configuration under the Microsoft SVVP Program. Should that stop all customers from using it? Consider the following facts:
- We have tested the configuration ourselves and haven’t seen any technical issues from powering on the DAG node VM automatically.
- Customers are using this configuration in production with no technical issues.
- Microsoft customers with a Premier Support agreement are entitled to reasonable support beyond specific SVVP-validated configurations.
In some cases customer needs can go beyond the vendor’s official support positions or guidelines. In these situations, it’s up to the vendors to step up and meet those needs. We’re extending an open offer to the Microsoft Exchange team to help them better understand, test, validate, and support this valuable solution!
Alex Fontana and Scott Salyer
(Original authors of the Exchange on VMware Best Practices and availability docs)