Testdriving the Azure-to-Azure (A2A) variant of the Azure Site Recovery (ASR) services


About two weeks ago the public preview was announced for an Azure-to-Azure site recovery service (DR). This was a highly anticipated service to me. So I’m glad it was announced! Today’s post will be on doing a quick setup of the service to get an insight into the look & feel.



What will we be doing today? We’re going to replicate a machine from “West Europe” to “South UK”.

What will the flow be? We’ll have a “Site Recovery Vault” in the target destination. Why? This to ensure that we can leverage the service when the source region went offline. In the source region, we’ll have our VM + Storage Account / Disks, but also a storage account used for caching. All changes make to the disk, will be stored in that cache. On the target side, we’ll have a storage account where a replica (and retention points) of the VM will be kept. Sound pretty easy? Let’s find out…



Let’s go to the marketplace and type “oms” ;

Select “Backup and Site Recovery (OMS)”, and we’ll deploy that one to “UK South”.

Once deployed, go to “Getting Started” and “Site Recovery”. On this page, click on “Prepare Infrastracture”.

On this page, select “Azure” as both the source & destination.

And you’ll notice that we don’t need to prep any infrastructure… (compared to a typical On Premises to Azure setup!).

So let’s continue to “Step 1: Replicate Application”

Select the source…

Next up, select the VM(s) you want to replicate.

And next up, select the target location. Next up, you can customize the target resources…

Which kinda looks like…

We’ll leave that one as is. We can also take a look at the replication policy…

Which looks like… As I’m fine with that, I’ll leave it as-is.

Now let’s “Enable Replication”!

The “Deployment” is started. It’ll create all the supporting resources, install the agent onto the VM and get the replication started.

Bare with, as this will take some time…

You can click on “Enabling protection” to get the status details.

And after a while, you’ll see that your VM will get the “Protected”-status.

Now we can change the “Compute and Network” settings (if wanted).

Or take a look at the disk status.

All looks great. So let’s do a “Test failover”. Here I chose to select a custom recovery point…

Now I need to select the network to which I will do my test deploy.

Let’s press “Ok” and wait for it…

I kinda did a stupid move, and selected the ASR resource group for my test deployment. That currently contains my ASR related storage accounts & network.

Though let’s check up on the status… It’s completing the test failover. 

Let’s check on the status of that one.

And a few minutes later it finishes.

Now we have a newly created VM + NIC.

I’m going to add a public IP to this virtual machine. This to see if all went well… Just to be complete, in a normal situation, you’re going to automate / script this as part of your recovery plan!

I’ll add a DNS name to my public ip too…

And it seems my failover went fine!

This as it looks just the same as my source VM.

As I’m happy with my failover test, let’s clean up the resources.

Once we started that one, we’ll see the same kind of process kicking in…

Again we can see the details.

And we are done!

Common Issues

If a machine is grayed out on the selection, this means that the machine is not in a running state!


If you are very eager, you might see the following error message. This means that the VM is running, but the VM agent isn’t ready yet. Have some patience… 😉


After the cleanup, be sure to also cleanup the resources you created additionally… Like the public ip address!


Preview Constraints

Check the support matrix… At this time there are several “gotchas” ; windows 2016 support, managed disks support, limited amount of regions, … Though these will get fixed by the time this service will be “General Available”.


Closing Thoughts

Any Enterprise looking to leverage Azure needs to have a BCP/DRP. Azure Site Recovery services shows that it has the capabilities to support this. When test-driving this service, I must say the process was very smooth. As I’ve done several “hybrid” ASR deployments, I must say those cost a bit more time, despite even being quite easy in setup.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.