June 16, 2022

What’s the difference between a 2-node cluster and a Stretched cluster?

Stretched cluster and 2-node cluster are two extremely powerful vSAN configurations that help our customers solve different types of challenges. Since both topologies offer synchronous data replication on two different sites (vSAN Fault domains), we often get asked by customers what’s the difference between these two. In this blog post, we will take some time to cover some of the most important similarities and differences that these two configurations have between each other.

Stretched cluster and 2-node cluster are two extremely powerful vSAN configurations that help our customers solve different types of challenges. Since both topologies offer synchronous data replication on two different sites (vSAN Fault domains), we often get asked by customers what’s the difference between these two. In this blog post, we will take some time to cover some of the most important similarities and differences that these two configurations have between each other.

To begin with, a 2-node cluster is the smallest configuration vSAN has to offer since it only requires two data nodes and one witness node.  Therefore, this topology is the perfect solution when the customer’s key requirement is to have a small but resilient configuration at a remote location. 

Stretched cluster, on the other hand, is a way to stretch a normal vSAN cluster across two sites and have the VM data redundantly replicated on both sites, with a third witness site to store the witness components. Having the data synchronously written in two different places, allows customers to have minimal downtime in case of a site failure. In addition to the high level of data availability and minimal downtime, a stretched cluster topology also gives the option to scale up to 20 hosts per site, to a total of 40 nodes + 1 witness node.

Note that it is possible to deploy a stretched cluster with only two data nodes and one witness node on a third site, but then the option for a secondary level of data resilience will not be available. Meaning the VM’s object will have two replicas on both sites and no additional local protection. Keep in mind that the network requirements for a stretched cluster should also be taken into consideration in this scenario.

Now let’s dive deeper into the comparison between these two topologies!

2-node and Stretched cluster configuration

The VMware vSAN 2-node cluster is a configuration where both data nodes are on a single site. The 2-node vSAN network is connected to a switch, or via a Direct Connection. The vSAN witness host that provides quorum for the 2 Nodes can be hosted on a third site via low bandwidth/high latency links, or on alternate infrastructure at the same location.

2-node cluster

The vSAN Stretched Clusters topology refers to a deployment where a user sets up a vSAN cluster with 2 separate active/active sites with an identical number of ESXi hosts distributed evenly between the two sites. The sites are connected via a high bandwidth/low latency link. A third site, hosting the vSAN witness host, is connected to both active/active data sites. This connection can be via low bandwidth/high latency links. To clarify, when using vSAN Stretched Clusters in a single data center, different rooms, or different racks could be considered separate sites.

sc cluster

 

 

Network Requirements

The easiest way to spot the differences and the similarities in terms of network requirements is to put these in a table to easily compare them. Let’s look:

 

2-node cluster

Stretched cluster

vSAN communication between the data sites

VMware recommends that vSAN communication between the vSAN nodes be over L2.

Layer 2 or Layer 3, with the following considerations:

-       Stretched Layer 2 does not require static routing

-       Layer 3 requires static routing to communicate properly between sites.
 

vSAN communication between the data sites and the witness host.

VMware recommends that vSAN communication between the vSAN nodes and the Witness Host is

-       Layer 2 for configurations with the Witness Host in the same site

-       Layer 3 for configurations with the Witness Host in an alternate site.

Layer 3 

Witness traffic separation

Yes

Yes

RTT (Round Trip Time) max latency between the data sites

5ms

5ms

RTT (Round Trip Time) max latency between the data sites and the witness host

<500ms latency RTT

<200ms latency RTT (up to 10 hosts per site)
<100ms latency RTT (11 to 15 hosts per site)
<500ms latency RTT (1 host per site)

Bandwidth between the data sites

Please refer to the vSAN Stretched Cluster Bandwidth sizer

 Please refer to the vSAN Stretched Cluster Bandwidth sizer

Bandwidth between the data sites and the witness host

Please refer to the vSAN Stretched Cluster Bandwidth sizer

Please refer to the vSAN Stretched Cluster Bandwidth sizer

Common features for 2-node cluster and Stretched cluster

Let’s also take a look at some of the vSAN features that are available for both cluster configurations.

cluster configurations.

·        Gateway override

The default gateway override feature in vSAN 7 U1 helps make deployments easier with fewer configuration mistakes. By eliminating the need to create static routes for each host participating in 2-node and stretched clusters, it helps avoid one of the most common configuration challenges in these environments.

·        VLCM

Lifecycle management of 2-node clusters and stretched clusters is improved with vSAN 7 U3 in that the virtual witness host appliance can now be managed and updated using vSphere Lifecycle Manager (vLCM).  VLCM will make sure to coordinate the update of the cluster in a proper way.

Note that witness host appliances shared across more than one 2-node cluster cannot be managed by vLCM.  In those cases, the witness host appliance must still be updated using VUM or replaced with a new installation of the witness host appliance.

Here, I would like to emphasize that the shared witness feature is only available for 2-node cluster configurations, and it is still not available for stretched clusters.

·        Improved resilience during failure of site and witness host appliance

2-node clusters and stretched clusters are sharing another common feature in vSAN 7 U3. With this version, both topologies are providing better availability during planned or unplanned events.  If a host is taken offline for maintenance (or fails), vSAN will use Adaptive Quorum Control (AQC) to recalculate the rules of quorum on an object-by-object basis.  Once it completes (which may take anywhere from several seconds to a few minutes), the data on the host will remain available and accessible if a subsequent maintenance event or outage occurs with the witness host appliance.

·        Second level of resilience

The Stretched cluster topology offers a second level of resilience on a per local site basis. This means that the vSAN stretched cluster is ensured against an entire site failure and against host failures within one of the sites. The 2-node cluster also offers a second level of resilience, but on a per disk group level. Thus, a 2-node cluster can ensure the VM’s data against a host failure and against disk group failure. This second level of availability for the 2-node cluster is the new nested fault domain feature introduced with vSAN 7U3. This level of resilience is not available for versions previous to vSAN 7 U3.

Licenses

A Stretched cluster configuration would require a vSAN Enterprise License, while a 2-node vSAN deployment model is not restricted to a specific vSAN license edition. In other words, any of the licensing editions can be used with a 2-node configuration.

Summary

2-node and Stretched cluster topologies have a lot in common and they also have their specific requirements in terms of licensing, networking, and configuration. For a more detailed overview of both configurations and their specifics, please check the following useful guides:

1.    vSAN Stretched cluster guide

2.    vSAN 2-node cluster guide

3.   Stretched Cluster Bandwidth sizer

4. vSAN Interactive infographic 

 

Filter Tags

Storage vSAN vSAN 2 Node vSAN Stretched Cluster Blog Advanced