If you have been paying attention to our blogs, you’ve heard us preach about the importance of enabling continuous workload optimization in your environments. You’ve learned about Business and Operational intent, a main component of your self-driving datacenter, which allows you to configure vRealize Operations to meet your specific datacenter goals and purpose. You’ve seen how optimizing workloads can help you save money on licensing costs, drive performance and meet SLAs, reclaim and repurpose unused hardware, and ensure you workloads are running on compliant infrastructure. All good things!
Over the last few releases of vRealize Operations we have made huge improvements to our workload optimization feature set by adding in host-based placement, scheduling and automation, simplified configuration workflows, tag violations, and much more. Well the train keeps rolling in vRealize Operations 7.5 as we add three key features: vSAN Support, Host Group Optimization, and Improved Tag Violation Visualizations.
Let’s start with the first one, vSAN Support. vSAN will be an important component for you as you build out your Software Defined Datacenter (SDDC). Because it is important to you, it needed to be supported by vRealize Operations, so with 7.5 you can run Workload Optimization on vSAN clusters. But in order to do it right we needed to take some of the nuances of vSAN into consideration so vRealize Operations 7.5 is Resync aware, respects vSAN Slack Space, and honors Storage Policies. Let’s take a quick look at each of these.
What is a vSAN resync? Simply put it’s the creation, movement, or repair of data to ensure assigned levels of resilience. vSAN must ensure availability of data under a variety of planned and unplanned conditions, it automatically manages the placement of data in accordance to the assigned storage policies of a VM. This placement of data is an ongoing operation that occur at any time vSAN determines the need to do so. This creates “resynchronization traffic.” This is simply I/O processing that is not being generated by the guest VM, but rather, the system, in order to meet the performance and resilience settings as prescribed by the storage policy. This can be triggered as a result of object policy changes, host or disk group evacuations, or object and component repairs. vRealize Operations will not generate a workload optimization plan if any vSAN clusters are currently running a resynchronization, in order to not get in the way during this time. Simple.
Respects Slack Space
Since vSAN is a distributed storage solution, it needs free space to perform certain actions that are transparent to the user. This free space is known as Slack Space. Any number of changes (e.g. changing a policy, host/disk group evacuations, on disk format changes) will require the use of this slack space temporarily to adjust to the new conditions. vRealize Operations is aware of this need and will not make any workload optimizations recommendations that may infringe on the slack space requirements of vSAN.
Honors Storage Policy
Storage Policies are an incredibly powerful tool that allow you to define storage protection and performance requirements using a set of rules, that are applied prescriptively to VMs and VMDKs. It allows you to easily set Failures to Tolerate (FTT) at various levels to meet demands of business. Simply define the outcome in the storage policy, then apply it to the VM. It’s that easy. vRealize Workload Optimization will NOT move a VM if it breaks the Storage Policy of that VM or VMDK in the process. It leverages vCenter and Storage vMotion to verify the target location will properly support the storage policies before making any moves.
Host Group Optimization
The next improvement comes into play when you are using the host-based placement feature of Workload Optimization. As you may recall, host-based placement allows you to drive your business intent down to DRS where vRealize Operations automatically created VM groups, host groups and affinity rules for you to meet your business objectives (e.g. license enforcement). Previous versions gave you visibility into the cluster demand for CPU and memory. With 7.5 you can also see CPU and memory demands on the host groups themselves.
That’s nice, but we didn’t stop there. We also added host group optimization which will alleviate stress on host groups when they are overloaded and running out of resources. This means if a host group, like the one above, goes non-green due to the high level of resource requests vRealize Operations will set the Not Optimized flag and will automatically move some of the VMs to another cluster with a similarly tagged host group. Ensuring your application are performing well all while meeting your business intent.
Let’s look at a simple example. We start with an environment on Day 0 where we want to drive host-based license enforcement. vRealize Operations and DRS have created the host groups and affinity rules to ensure we are meeting our business goals. As we add workloads to the environment they are running on the appropriate hosts with DRS and vRealize Operations maintaining performance balance.
As more and more workloads are added (note the workloads with asterisks) we start to see a performance hit on Host 8 which is a part of the MSFT host group.
DRS is unable to resolve this resource contention as there are no other hosts on this cluster that are valid for these workloads. vRealize Operations Workload Optimization will automatically run a workload plan and move some of the workload from Host 8 to Host 4 on Cluster 1 thereby relieving the stress on the VMs and ensuring we are still in meeting our licensing restrictions.
Improved Tag Violation Visualizations
The last improvement is my favorite one because I have already seen the benefits from it in my work lab about a dozen times or more. Its pretty simple, but beneficial. In previous versions vRealize Operations could flag a datacenter as Not Optimized if it wasn’t meeting your Business Intent (see below). It would show you the vCenter tags that were not being honored, but not much more. You had to start looking around as to what cluster, host and VM was causing the trouble. Kind of tedious.
Now, in 7.5, it shows you the tag being violated, the cluster its associated with, AND the VM(s) that is causing the violation! A huge time saver.
Just give in and download a trial of vRealize Operations 7.5 and try it out! You can also find more demos and videos on vrealize.vmware.com. Explore and find out for yourself how Workload Optimization has improved in 7.5!