3 Azure VM management concepts to prevent future headaches

Theresa Miller explains how to avoid pitfalls from IP address assignments, backups, and high availability.

Whether new to Microsoft Azure or not, there are three key fundamentals that if missed, can cause you pain points within your virtual server deployment. Just missing some of these steps can potentially cause a critical outage or leave your virtual machines in a state where recovery cannot occur without a full server rebuild. These key Azure VM management considerations fall into the realm of Azure default IP address assignments, backups, and high availability.

Let’s take a look!

1. Azure default IP address assignments

When first deploying a server in Azure, an IP address will be assigned to your virtual machine (server). If you look in the Windows operating system (OS) network card settings you will not see a hard-coded IP address. Everything for IPs will be set to automatically be configured. This essentially is acting as a DHCP- assigned IP address. In the Azure console, Microsoft will show you the IP address, or you can find it using the traditional ipconfig/all from a command prompt within the server. At first glance you may think all is OK, but the main takeaway is that it is a DHCP-assigned IP address. This means that if the server is rebooted, a new IP will be picked up.

This is a big problem, and one that I ran into recently when supporting a customer. The customer thought they were good to go and had even locked these IPs into external DNS to launch their web-based application. Well, after several months, they decided to run Windows updates. Guess what—when the server was rebooted, a new IP address was assigned to the server. The web application was now broken, and all the external DNS addresses for the application were now bad. Then, to top it off, the server wouldn’t boot. (I will come back to this Azure VM management issue later.)

So, now the customer has a server that can no longer get its IP address (when that IP is gone, it’s gone) and it won’t boot. The customer didn’t know that within the Azure console you can change a setting allowing DHCP-assigned IP addresses to become statically assigned forever. In the network card properties within the virtual Windows server OS, the IP address still will not be hard coded like we are used to in the enterprise, but at least this cloud-hosted virtual server will function as expected and not lose its IP address upon reboot.

2. Backups

By default, your Azure-based virtual machine will not be backed up. So, for an easier system recovery, this is a simple step that can be taken right from within the virtual machine’s Azure configuration. One that could save your administrator major headaches and stress if a system suddenly becomes unable to reboot properly.

Consider the scenario above, where a reboot after patching left the system unusable. In the customer case above a support ticket with Microsoft was needed, because there weren’t any backups of the server. Thankfully, Microsoft had a way to repair the server, but it took a lot of time. The system was unavailable until the next business day. This particular customer could wait a day, but not every business can afford that option.

To initiate a backup of your server from within Azure, go to your virtual machine. Under the left-side Operations section there will be an option for backups, where you can start the process and schedule it to occur routinely. Note: Microsoft changes the console often, so at some point these exacts steps may vary slightly. This simple Azure VM management step could have allowed the customer to more quickly recover their server, and be back in business much sooner.

3. High availability with Availability Sets

While we are covering Azure backups, it’s important to also look at high availability. This option is another that will not be applied by default, and cannot be applied after the virtual machine is built. Microsoft recommends recreating the machine if you find yourself in this situation.

If your expectation is that a full datacenter outage will not impact your virtual machine, then you need to plan to use Azure Availability zones from the beginning.

  • First, be sure that your servers are located in a region where Availability Zones are available, and that your subscription will also support your ability to use those regions. If this isn’t in place then you won’t get very far.
  • Then, Availability Sets can be configured to finish the configuration and deploy the level of high availability you are looking for.

Concluding Thoughts

Today, we really only scratched the surface on each of these Azure VM management topics. Beginning to understand how these key concepts work before you start working with Azure will go a long way for ensuring that you have the robust virtual machine deployment you are expecting. Eliminating those unexpected surprises that can occur when these items are not configured correctly up front.

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.