How to enable High Availability (HA) on the network between two availability zones
In this tutorial you will learn how to enable High Availability (HA) on the network to add redundancy between the two Madrid availability zones and avoid the network service depending on a single zone.
By the end, you will have a network with HA configured between both zones and you will know how to check that the configuration and connectivity between your servers are correct.
👍 What you get
A network with High Availability enabled between the two availability zones, ready to be part of architectures that require continuity in the event of a zone outage.
Before you begin: when you need network HA
Not all architectures that use two availability zones need to maintain a network with HA.
If you have components permanently deployed in both zones and also need to add redundancy to the network service, enabling HA prevents routing from depending on a single availability zone.
On the other hand, if you use the second zone mainly as a recovery target (for example, to deploy a server and restore a backup when the original zone is unavailable), you do not necessarily need to keep network HA enabled. The configuration will depend on the continuity scenario you have designed for your customer.
What network HA does and does not do
The same network can be used from servers deployed in MAD1 (or DOJA in the image) and MAD2 (or DOJO in the image) even if HA is not enabled. Connectivity between zones does not depend on this setting.
When enabling High Availability, Plenit duplicates the virtual router that manages the network and deploys an instance in each availability zone. If the router in one zone becomes unavailable, the one in the other zone takes over without interrupting traffic.
Networks Standard and VPC allow High Availability to be enabled. Networks L2 do not have this option because they do not have their own router. Network services are managed by the partner, usually through a VNF such as FortiGate or WatchGuard.
Prerequisites
To follow this tutorial you need:
At least two operational servers, one in each availability zone.
Both servers connected to the same Standard or VPC network.
If you followed the tutorial How to deploy servers in multiple availability zones, you already have this setup prepared.
Step 1. Access the network configuration
From the Servers, go to Networks/IPs.
You will see the list of networks in the subscription and the status of High Availability for each one.
Locate the network that connects the servers deployed in the two availability zones.
Step 2. Enable High Availability
Turn on the High Availability toggle for the network.
The platform will ask for confirmation before applying the change.
Confirm the operation and wait for the platform to complete the configuration.
Step 3. Check the configuration
Once the process is complete, check in the network list that:
High Availability appears enabled.
The network location shows the two Madrid availability zones: DOJA and DOJO, corresponding to MAD1 and MAD2.
With this setup, the network's virtual router has an instance in each zone.
Step 4. Verify connectivity
Check that there is communication between the servers deployed in the two zones.
You can use ping to perform a basic check:
Perform the test from a server in DOJA (MAD1) to the server in DOJO (MAD2) and check that you receive a response.
If you use Windows Server, keep in mind that Windows Firewall may block incoming ICMP traffic by default. Therefore, a ping failed one does not necessarily mean there is no connectivity between the servers.
Step 5. Validate the environment before production
Before deploying critical workloads, check that the infrastructure matches the intended design:
The servers are deployed in the correct availability zones.
Both use the intended network.
The network has High Availability enabled.
There is connectivity between the components that need to communicate.
Use the diagram view in the Servers section to visually check the distribution of the servers and their networks.
Connectivity between the two availability zones is designed to offer low latency, with values of around 1 ms and always below 5 ms RTT. If you are going to deploy applications that are especially sensitive to latency, also check the actual latency between their components and compare it with the application’s specific requirements before going into production.
Conclusion
You already have a network with High Availability between the two Madrid availability zones. The routing service has redundancy between DOJA and DOJO (MAD1 and MAD2) and does not depend on a single router located in one zone.
Network HA is part of the continuity strategy. Full service availability will also depend on how you have distributed and configured the servers, applications, and data.
Last updated
Was this helpful?

