It Should Be Easy to Upgrade Your Load Balancer

Applications have never been more important in business than they are today. And where there are applications, there’s a load balancer, working behind the scenes to ensure your applications can be used comfortably and safely at all times. When operating a load balancer, the most troublesome issue is upgrade work. Let’s examine the problems of traditional load balancer upgrades and take a look at VMware’s automated, streamlined solution: NSX Advanced Load Balancer.

Why do you need to upgrade your load balancer in the first place?

The main reasons to upgrade a load balancer are to patch vulnerabilities and bugs, to enable new features, and for EoSL support. A load balancer, located between users and applications, must above all be stable; we particularly want to avoid service disruptions due to defects in the load balancer software. For this reason, load balancer upgrades are inevitable. And for IT, the trick is to make them, as transparent and painless as possible..

For illustration, let’s take a look at a recent international case involving load balancers. To address software glitches, a project was running to upgrade hundreds of load balancers. The upgrade work was carried out little by little over several months, with operations personnel setting up maintenance windows and devoting a lot of time to the effort. But just as the project was about to end, a new software vulnerability was discovered — and those hundreds of upgrades were needed all over again. As you can see, finding ways to streamline the upgrade process can have a big impact on the time and productivity of operations personnel.

Problems with traditional upgrade methods

Now let’s consider the problems with traditional upgrade methods.

Since many operations are performed manually by the operator, upgrades take time and are prone to mistakes.With a conventional load balancer, based on an Active-Standby configuration, you need to (1) upgrade the Standby machine, (2) switch, (3) upgrade the old Active machine, and (4) switch back. At each step, it’s necessary to carefully check the status of the equipment. The procedure puts a heavy burden on the operator and is prone to work mistakes.

When operating multiple load balancers, it takes time and effort to work one by one.It’s not uncommon for load balancers to operate on multiple systems. Furthermore, it’s increasingly common to operate not only on-premises but also in the public cloud. The greater the number of load balancers you manage, the greater the time and effort it will take to upgrade.

Service impact during upgrade workDuring the upgrade process, Active-Standby devices will be switched. At that time, the existing session is disconnected, so a service impact is unavoidable. You need to pre-arrange the date and time of work with the person responsible for the application or LOB.

NSX Advanced Load Balancer’s innovative upgrade process

VMware enables customers to streamline upgrades and save operations personnel time and effort with the NSX Advanced Load Balancer. Specifically, VMware’s load balancer solves the challenges of traditional upgrade methods by:

Automating most of the upgrade process
Upgrading multiple load balancers at once
Eliminating service impact during upgrade as much as possible

The secret to achieving a streamlined upgrade capability lies in the load balancer architecture. The NSX Advanced Load Balancer is a multi-cloud, 100% software load balancer with separate control and data planes. The control plane (Controller) provides integrated management, analysis visualization, and orchestration. The data plane (Service Engine) is the entity of the load balancer and provides functions such as server load balancing, GSLB, and WAF. The advantage of this architecture is that you can centrally control all load balancers from a single controller. You also have the flexibility to scale out and scale in your load balancer to better control traffic and minimize service impact during the upgrade.

Simple steps to upgrade NSX Advanced Load Balancer

Let’s take a step-by-step look at how to upgrade your NSX Advanced Load Balancer. This time, for the sake of clarity, I will introduce the method using the GUI. Of course, you can do the same with the CLI or API.

[Upgrade from Version 20.1.2 to 20.1.3]

Upgrade Load Balancer from Version 20.1.2 to 20.1.3

Upload the image file to the controller
After logging in to the controller, go to Administration – Controller – Software and click Upload From Computer to upload the image file.

Confirm that you have uploaded the image file.
Perform upgrade
You can also specify the scope of the upgrade in detail. For example, you can upgrade only the controller, or just the load balancer group for a particular tenant or system. In this example, we want to upgrade the entire system (controller and all Service Engine Groups).Go to Administration – Controller – System Update, select the image file, then click UPGRADE.

If desired, you can specify the Service Engine Group to upgrade and then change behavior if the upgrade fails. In this example, we will use the default settings. Click Continue.

First, start upgrading the controller.

When the controller upgrade is complete, we then upgrade the service engine, which is the entity of the load balancer. We can perform rolling update upgrades on a per Service Engine Group basis. For example, if two Service Engines are running in the same Service Engine Group, upgrade them one by one. The Service Engine being upgraded will stop service, while the other Service Engine will continue service.

Execute the following command on the management PC to check status during the upgrade:
- Ping Service Engine1 (172.16.130.131)
- Ping Service Engine2 (172.16.130.132)
- Execute httping for HTTP Virtual Service (VIP: 172.16.130.104)
After a short time…

At the end of all Service Engine upgrades, the number of failed httpings for the HTTP Virtual Service VIP is zero.

Make sure that the controller upgrade completes successfully.

Make sure that the Service Engine Group upgrade also completes successfully.

And that’s it. Upgrading NSX Advanced Load Balancer is that simple.

Summary

With NSX Advanced Load Balancer, even if hundreds of load balancers are deployed across multiple data centers and public clouds, all upgrade work can be done with just two steps: (1) upload the image file to the controller and (2) execute the upgrade. A simple, automated process that can dramatically streamline the operation of the load balancer.

Learn more about NSX Advanced Load Balancer.

Interested in testing out our various products on the cloud? Try our Hands-on Labs.

[Supplement 1] About rollback method

If something goes wrong after the upgrade is complete, you can quickly and easily roll back to the original version. You can also specify the rollback target range in detail. For example, if a defect is found only in a specific service, select an image that rolls back only that Service Engine Group.