High Availability Admission Control, Part 1
VMware has announced the upcoming vSphere release 6.5 at VMworld Europe Barcelona in the middle of October as part of its new cross-cloud-architecture. Apart from VMwares strategical plans around vSphere 6.5, vSAN 6.5 and vRA 7.2, I will discuss some of the new features in vSphere 6.5 from a deep technical view yet and in following parts of this article. I´ll start with vSphere HA.
There a lot of new features in the VMwares HA-Implementation for vSphere (fdm) and the corresponding settings in web vCenter, to configure them.
vCenter Server High Availability
vCenter Server 6.5 provide a new feature called vCenter Server High Availability, which replaces the vCenter heartbeat solution available in older versions. The feature itself will be only available on the appliance.
This solution follows a concept with active-, passive- and witness-nodes – all of them are cloned from the existing vCenter Server. I will write more about “vCenter Server High Availability” in an additional blog post later.
VMware also introduced changes within the vSphere HA module
Admission control has been completely overhauled from a GUI-point-of-view, so it becomes much easier to use.
Within the GUI, administrators can now see a link called “Admission Control” as a sub-menu within “vSphere HA” on the left hand side of the cluster-settings-dialog. In addition, the admission-control-policies are divided into three sections: This three sections are “Host failures cluster tolerates”, “Define host failover capacity by” and “Performance degradation VMs tolerate”.
The default-policy of “Define host failover capacity by” is “Cluster resource percentage”, but you can choose either “Disabled”, “Slot-Policy” or “Dedicated Failober Host” from the underlying menu. The first and the third sections only have numeric values such as the number of hosts in the first case and a percentage value in the third case. This results in the well-known default-behavior of admission control: If you have a two-host-cluster, the value for “Define host failover capacity by” is set by 50 % and the “Host failures cluster tolerates”-Setting can only ne “1”.
Never waste any resources again
But, the failover-capacity of 50 % is automated calculated depending on the number of hosts in cluster and the first policy. The advantage is, when you add a host to the cluster, the amount of resources will be automatically changed. vSphere HA will adjust the percentage-value automatically depending on the selected number of failures you want to tolerate and the amount of host within the cluster. This solves an old problem: Wasting resources, when administrators forget to adjust the percentage-setting after having added a new host to the cluster. And another great thing: The new behavior works without using “slots”.
In the following parts of this blog-series, I will discuss some further aspects of admission control, additional features of vSphere HA like restart priority enhancements, HA orchestratedRestart or pro-active HA and then a lot of other new feature in vSphere 6.5, for example from vSphere DRS, Virtual SAN or VVOL-Support.