Difference between HA and DRS in VMware vSphere

VMware vSphere Clustering Features: HA and DRS

HA and DRS are clustering features of the VMware vSphere platform meant to come in handy on different occasions to mitigate host failures and VM data transfer errors. A vSphere cluster refers to two, three, or more ESXi hosts connected together over a network – sharing the same hardware resources such as memory, processor, and storage on the host system.

You can manage VMware vSphere clusters in vCenter and ESXi hosts included in a cluster are called “cluster nodes.” That said, there are two types of VMware vSphere clusters: High Availability (HA) and Distributed Resource Scheduler (DRS). To deploy these vSphere clustering features, the following conditions must be met.

• You must have two or more ESXi hosts with the same configurations such as the same processor family (Intel or AMD), patch level, and ESXi version. However, it is advisable to use at least three ESXi hosts for more effectiveness.
• You need high-speed network connections for all networks used: management network, vMotion network, and storage network. Also, you will need redundant network connections.
• All clustered ESXi hosts must share the same datastore or have access to one particular datastore. You can use a Network Attached Storage (NAS), Storage Area Network (SAN), or VMware vSAN as the shared datastore.
• Your connections should support NFS and iSCSI protocols for access to the data on the shared datastore that stores the VMs’ files.
• Your installed ESXi host versions must be compatible with VMware vCenter Server.

Note: No quorum is needed for VMware vSphere cluster unlike in a Hyper-V Failover Cluster.

What is VMware HA?

VMware High Availability (HA) is a vSphere clustering feature that automatically restarts a virtual machine (VM) in case of failure (such as experiencing a failed ESXi host). This allows the VM and its connected applications to continue running provided they’re all in the same vSphere cluster. So, with VMware HA enabled, even if an ESXi host fails, the VMs on the host would be automatically on a healthy host so the end users will only experience a short downtime.

vSphere HA Requirements

To run vSphere HA in a clustered environment, your virtualization environment must meet the following requirements.

• VMware vSphere Standard license
• At least two ESXi hosts with the same configuration
• The hosts must be using a shared datastore and utilize at least 4GB RAM per host.
• High-speed network connections for all networks used and redundant physical network connections
• A pingable gateway
• Your ESXi host versions must be compatible with VMware vCenter Server

How does vSphere HA Work?

It simply checks your ESXi hosts looking out for failures; once a host failure is detected, vSphere HA will automatically migrate and register the VMs on the failed host to a healthy one in the same clustered environment, then restart the VMs. You can still access the VM files such as the VMX and VMDK files on the shared datastore – vSphere HA does not migrate the VM files.

Apparently, there will be some downtime during the migration process; the length of the downtime depends on the amount of time needed to restart the migrated VM(s) on another host. More so, VMware HA does not rely on the guest operating system installed on the VM. When vSphere detects a failed host, only the CPU, memory, and network components of the VMs deployed on the failed host are changed to that of the healthy host after the migration.

What is a VMware Cluster?

A VMware cluster is an environment of multiple running ESXi servers. ESXi allows creating and running multiple VMs on a single server, but when you’ve got more than one ESXi server or host, you can form a cluster by connecting to vCenter. VMware vCenter allows you to run and manage multiple clusters. To form a VMware Cluster, you have to use the same server configuration for all the ESXi hosts you want to include in the cluster.

vSphere Cluster requirements

You must use the same server environment for all ESXi hosts in the cluster, most especially, the same CPU. If one of your hosts is using AMD processors, and the rest are built on Intel processors, only the hosts on Intel processors can be used in the same cluster.

Only when you have set up a VMware vSphere Cluster environment can you implement HA or DRS, or both at the same time.

How to Configure HA and DRS in VMware?

vSphere HA

• Log in to VMware vSphere Client and go to “cluster” in the Hosts and Clusters menu.
• Right-click on the cluster you want to edit and select “Settings.”
• Go to Configure 🡺 vSphere Availability 🡺 Edit 🡺 toggle the vSphere HA switch
• Go to the Failures and Responses tab and activate Enable Host Monitoring.
• You can manually edit the vSphere HA configurations in the Advanced Options tab.

Save your edits – click OK.

vSphere DRS?

• In VMware vSphere Client, go to the clusters page, click on the “Configure” tab, further click on vSphere DRS in the Services pane, and click EDIT.
• Toggle the vSphere DRS switch
• Go to the “Automation” tab and choose between Fully and Partially automated options. Also set Migration Threshold.
• Get into the “Additional Options” tab and configure VM Distribution, Scalable Shares, and CPU Over-Commitment.
• Finally, tune the Distributed Power Management settings and make any further adjustments in the “Advanced options” tab.
• Click OK to save and exit when done.

Using VMware HA and DRS Together

Using VMware High Availability (HA) and Distributed Resource Scheduler (DRS) together enhances the availability and efficiency of your virtual environment. VMware HA provides rapid recovery from outages by automatically restarting virtual machines (VMs) on other hosts if a server fails. DRS, on the other hand, optimizes VM performance by distributing them across hosts in a cluster based on resource availability and predefined rules.

When you combine VMware HA with DRS, you get a robust system that not only ensures your VMs are always running but also that they are running optimally. DRS can even respond proactively to avoid potential issues by moving VMs before a host becomes overloaded or encounters a problem, reducing the likelihood of VM failure and the need for HA intervention.

In this integrated setup:

• HA monitors the health of physical servers: If a host fails, HA steps in to restart its VMs on other hosts within the resource pool, minimizing downtime.
• DRS optimizes resource utilization: It continuously monitors resource usage across hosts and VMs, balancing loads to prevent any single host from becoming a bottleneck, which in turn helps in maintaining a more stable and reliable environment suitable for HA operations.
• Enhanced coordination: When HA triggers a failover and restarts VMs on new hosts, DRS can immediately rebalance these VMs across the cluster to ensure optimal performance, preventing any one host from being overwhelmed post-failover.
• Preventive actions: DRS can also take preemptive actions based on predictive data about host resource usage, moving VMs before a potential host failure occurs, which dovetails with HA's goal of minimizing service interruption.

Using VMware HA and DRS together thus provides a dual layer of protection and efficiency, ensuring not only that your applications remain available but also that they perform well, adapting dynamically to changing conditions in the virtual environment.