Sergey Nivens - Fotolia
Combine vSphere's Distributed Resource Scheduler and High Availability tools to best ensure VM high availability.
By itself, VMware vSphere High Availability (HA) emphasizes VM high availability, but Distributed Resource Scheduler (DRS) can supplement it with affinity rules that take into account other considerations.
VSphere HA uses criteria to find the first available host system capable of restarting an afflicted VM, such as a system or resource, without considering the overall loading of those systems. The net result is that some systems receive heavier loads than others. Your first priority is to get all of the VMs running again.
VSphere can sometimes experience problems restarting VMs because of resource limitations. For example, features such as Distributed Power Management try to consolidate VMs onto fewer hosts and place the unused hosts in standby mode, which leaves the host systems unavailable for failover. Similarly, the resources needed to restart a VM might be available in the environment, but they could be scattered across multiple host systems, which also makes failover impossible.
VM high availability is best with vSphere HA and DRS
If you properly configure vSphere HA and DRS, the combination usually results in more evenly loaded HA clusters -- especially after failovers and restarts. For example, DRS can facilitate VM high availability by bringing hosts out of standby mode so that vSphere HA can perform a failover or by shifting VMs so enough resources are free for a VM failover.
DRS and vSphere HA also support VM affinity rules. You can choose how to enforce those rules to best support VM high availability via vSphere HA. For example, DRS and vSphere HA can both stipulate VM affinity rules that place VMs on particular hosts during failovers or anti-affinity rules that prevent VMs from failing over to certain hosts.
The trickiest part of this VM high availability technique is when you create or edit DRS affinity rules. You must choose whether vSphere HA respects DRS affinity rules. You can tell vSphere HA to respect the DRS VM anti-affinity and VM affinity rules during a failover event.
If vSphere HA respects DRS VM anti-affinity rules during a failover, vSphere HA won't fail over a VM to specific hosts. If vSphere HA respects DRS VM affinity rules during a failover, vSphere HA will attempt to place a VM on the desired hosts -- as long as the other failover criteria allow it.
Dig Deeper on Disaster recovery, failover and high availability for virtual servers
Related Q&A from Stephen J. Bigelow
Just because software passes functional tests doesn't mean it works. Dig into stress, load, endurance and other performance tests, and their ... Continue Reading
Don't neglect form factor as part of your data center server selection. Instead, figure out what type of environment you need and learn which server ... Continue Reading
Learn how load balancing in the cloud differs from a traditional network traffic distribution, and explore the different services available from AWS,... Continue Reading