everythingpossible - Fotolia

Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

VMware makes welcome changes in vSphere 6.5

VMware has rolled out a lot of new features in the latest version of vSphere, including updates to High Availability and DRS, as well as a new version of VVOLs.

The latest version of vSphere focuses heavily on simplifying the user experience and improving security with a built-in security system. But there are a lot of other new features encompassed in vSphere 6.5, including changes to High Availability, improvements to the Distributed Resource Scheduler feature and additions to Virtual Volumes.

Let's take a more in-depth look at these features, how they work and what they do.

VMware gives HA a face-lift

VMware released a new version of its virtual machine hardware, version 13, which increases the maximum configurations of VMs and supports new features like Secure Boot.

VSphere 6.5 also includes an enhanced version of High Availability (HA), which VMware has renamed vSphere Availability. The vendor tweaked how HA works, starting with tying failover capacity with host failures to tolerate in Admission Control. For example, in a four-node cluster, setting "host failures to tolerate" to two means the failover capacity is automatically set to 50%; if you were to set the host failures to one, the capacity would be set at 25%. Users have the option to override this setting if they so choose. 

VMware has also made changes to the Admission Control policy, which now defaults to cluster resource percentage, rather than slots. This is a big improvement, because it can be difficult to calculate the slot policy based on powered-on VMs. Since these settings are now configured by default, if you increase or decrease the number of hosts in your cluster, your failover capacity will update automatically to reflect your cluster resources.

In the past, HA would restart VMs with no order dependencies, which was a pain for multi-tier applications, or even for Active Directory authentication or domain name system if your virtual data center didn't start first. HA Orchestrated Restarts is a new feature that lets you define the order in which a specific multi-tier app needs to boot, such as the database server, then the app server and, finally, the web tier. Every time HA needs to restart this tier, it will do so according to your rules.

VSphere 6.5 brings major changes to DRS

One new feature in vSphere 6.5 has a slightly misleading name. Proactive HA is not an HA feature, as its name would suggest. Instead, it's a Distributed Resource Scheduler (DRS) feature that vacates a host as soon as it receives a notification of the host being in degraded mode, such as from a hardware management tool like Dell OpenManage or Hewlett Packard Enterprise Systems Insight Manager.

Once a host is in degraded mode, Proactive HA puts it into Quarantine Mode. This means DRS will start to evacuate VMs from the host -- assuming it doesn't affect performance of the cluster or violate any DRS rules -- and ensure VMs are not migrated to it unless absolutely necessary, such as when there aren't enough available resources in the cluster. Unlike Maintenance Mode, where the host is unusable, Quarantine Mode will still allow hosts to run VMs.

VMware has also made improvements to DRS, adding new options for rebalancing your cluster, such as VM Distribution and CPU over-commitment. DRS is now network-aware, too, which means it takes network utilization into consideration when using vMotion to move VMs between hosts.

Predictive DRS is another new feature in vSphere 6.5, one that won't be supported until VMware releases the next version of vRealize Operations.

Predictive DRS is another new feature in vSphere 6.5, one that won't be supported until VMware releases the next version of vRealize Operations (vROps). VROps will collect historical data, such as CPU and memory utilization, and use its analytics engine to determine whether a VM should be migrated based on a forecasted trend. DRS then acts on these forecasted trends, up to 60 minutes in advance.

This is useful for cyclical workloads where vROps can predict what could happen, such as a SQL server that undergoes high utilization at the end of the month when a report is regularly run.

The latest version of Virtual Volumes now comes with replication -- per object or replication group. And VMware debuted vStorage APIs for Storage Awareness 3.0 in vSphere 6.5 for better Storage Policy-Based Management integration.

It looks like vSphere 6.5 also has a new version of Virtual Machine File System (VMFS) that comes with support for 4 KB Native drives (4Kn) in 512e mode. It should be noted, however, the current implementation emulates 512 sectors on a 4Kn drive and not fully 4Kn. VMFS 6 includes the intriguing Automatic Space Reclamation feature, which uses vStorage APIs for Array Integration (VAAI) UNMAP commands. You may recall VMware tried to implement this feature in vSphere 5.0, but it caused problems with customers -- so many, in fact, that the company asked customers to turn it off, and turned it off completely with 5.0 Patch 2.

VAAI UNMAP allows you to reclaim capacity on the storage array once you delete a VM from a thin-provisioned LUN; in the past, you had to run the UNMAP command in order to reclaim the blocks. This is now integrated into the user interface, which allows you to turn the feature on or off. The new tool reclaims blocks that have been freed within 12 hours.

Now that vSphere 6.5 is generally available, try testing out these and other features to see if they can make a difference in your infrastructure.

Next Steps

Does vSphere 6.5 live up to expectations?

VVOLs promise a storage revolution for VMs

Fine-tuning vSphere High Availability

Dig Deeper on VMware new releases and updates