Step-By-Step: Azure Site Recovery – Controlled Failover

Hello folks,

Lately I wrote a few articles about protecting your physical servers with Azure Site Recovery or ASR for short.linkedin

The reason I have been really looking into this as I mentioned in the first post is because disasters have been occurring and I know the people that are involved.  In his post, Dave Kawula a Canadian MVP explained that he help the Local School board protect their workloads.  But I don't know if they ever had to actually pull the plug and initiate the failover.

Another of my friends was not so lucky.  In his case he did NOT have a proper DR site setup and when the datacenter flooded, they were out of commission for a long time.  They did have a Plan, but it was not up-to-date, it was not tested, and no-one really knew what “the plan” really was.  So in essence they did not have a plan.

But last week in Vancouver during MVPDays, someone asked me about failing over multi-tier applications. “How do i deal with workloads that have multiple machines involved” he asked.

Valid question… How do you deal with workloads like SharePoint for example that would have a SQL server on one physical server, the Front-end on another and the application server on a third.  how do you ensure that you can recover all three without problems.

It’s quite easy.  Let’s look into it.

Step 1 – Setup your environment.

We will assume that you have read the first 3 posts and that you have an environment that is already protected and that the machines you want to protect as a “solution” are already replicated.

Step 2 – Enable Multi VM Consistency

The second step is very simple (it’s the beauty of ASR).  In the Azure Portal.  click Recovery Services in the left hand side menu, then select the Site Recovery Vault you created.


In the vault, navigate to the protected Items, select the protection group that includes your multi-tier solution.


and in the configuration tab,  enable Multi-VM Consistency.  (adjust the RPO Threshold, Recovery point retention, and Application consistent snapshot frequency as appropriate.)


After all the machines in that protection group are protected we will also be able to ensure that all machines in the group can be recovered to a data consistent state.

Step 3 - Planning the Failover

I have rarely seen any sysadmin or other IT professionals that did not have a very specific order in which they would stand up a solution.  something like,

  1. Backend Database server started first.
  2. Then, the application\middle ware server,
  3. and finally the frontend server to complete the solution.

for automated failovers….  how do you ensure to follow that process? That also is easy…

in the recovery vault, navigate to Recovery Plans and select/create a proper plan that covers all the machines included in the solution.


in that Recovery plan, you will be able to create groups and move the proper machines into them in order to be able to determine the deployment order.


you’ll also be able to insert manual processes, or automated scripts at any point in the plan.  This is the level of control you have over the failover process.

I hope this was useful,  Now i’m going to sleep.  because i know my datacenter is protected…



Pierre Roman

Comments (0)

Skip to main content