In this second part, we’ll explore how the types of disks used in a host or cluster can directly influence the behavior and construction of a vSAN ESA cluster.

In my last article, Asymmetry in vSAN ESA – Part 1: What You Need to Know, I explained what it means to have an asymmetrical cluster and its potential impact on failure tolerance.

Read-Intensive vs Mixed-Use

Before diving into details, it’s important to understand the concept of DWPD (Drive Writes Per Day. This is a metric used by hardware manufacturers to indicate how many times the total capacity of a disk can be written per day over the life of the device. Higher DWPD means greater endurance.

There are generally two main disk profiles:

• Read-Intensive (RI): Lower endurance, cheaper, and larger capacity — suited to read-dominant workloads.
• Mixed-Use (MU): Higher endurance, more robust for both reads and writes — more expensive.

Tipo	DWPD	Utilização	Caracteristicas
Read-Intensive (RI)	1-3	Suited to read-dominant workloads.	Lower endurance, cheaper, and larger capacity.
Mixed-Use (MU)	3-5	Environments with a high write rate.	Higher endurance, more robust for both reads and writes, more expensive.

Changes in vSAN ESA

Initially, vSAN ESA supported only Mixed-Use (MU) disks. This made sense because in ESA architecture there is no separate cache tier — all disks handle both reads and writes, requiring greater robustness.

However, VMware has recently certified certain vSAN ReadyNodes to support Read-Intensive (RI) disks as well. This makes vSAN ESA more accessible, since RI disks are cheaper and offer larger capacities.

Even though it’s technically possible to mix RI and MU disks in the same host or cluster, this configuration is not supported by VMware.

Conclusion

If you’re starting with vSAN ESA and want to avoid performance risk, use Mixed-Use (MU) disks.

If your environment has a predominantly read-heavy workload, you can consider Read-Intensive (RI) disks — only if you use vSAN ReadyNodes certified for it.

And finally, never mix RI and MU disks in the same host or cluster — not only is it unsupported, it creates asymmetries that can compromise the environment.