Have you ever wondered what is going on in your virtual infrastructure? Have you ever struggled to see when do you need to scale up or scale down your host infrastructure that runs your VMs? Have you ever thought how do you properly do the capacity planning of your virtualised infrastructure?
If your answer to any of the above is yes, then there is a strong case for you to start using VMware vCenter Operations Management Suite and it’s capacity planning capabilities that will save you from doing manual calculations and human errors while doing capacity planning, adding/modifying extra capacity and modelling that change manually. This post is intended for helping you to understand the “Planning” capability of vC Ops.
When you login to the vSphere Dashboard (default portal) in vC Ops you will see different “Worlds” in the left rail. When you expand the vCenter, Datacenter and select a Cluster, choose the “Planning” tab on the right side and you will see the Summary of the Cluster object.
This will show you a preliminary capacity dashboard, network I/O trending and forecasting. Also gives you flexibility to do What-If modelling of CPU & Memory. You have two options to do the initial planning. You can either choose Physical View or choose Virtual View.
You can also customize the aggregation type and perspectives as well to match your need for doing the infrastructure planning. Common Virtual Machine configuration that distributes the percentage of your VMs into different bucket is also possible to export in just one click.
However, the most powerful tool in the capacity planning area in vC Ops is using What-If analysis or What-If Modeling of your infrastructure needs. It lets you visualise what happens to your capacity when you add or remove virtual machines, add or update your physical hosts, as well as add or update your datastores. You can even choose the existing VM as your base model to craft the What-If Modelling or you can specify a model configuration at the run time as well.
Just keep it in mind that What-If Modeling can only be done on either Cluster or Hosts level. You are given three ways to view the trend and plan the capacity accordingly. You need to select a Cluster object from the left side of the tree, go to the Planning Tab and select “New What-If Scenario”. The three views correspond to the three views in the Time Remaining view gallery.
- Average Virtual Machine Capacity – Used
- Virtual Machine Capacity – Summary
- Cluster or Host Capacity Usage – Trend
If you have virtual machines those are close to the average virtual machine in the environment in terms of configuration and performance, then vC Ops can estimate future problems and capacity shortages more efficiently.
Remaining Physical CPU/Memory
Remaining VM in terms of CPU/Memory = Demand Average CPU/Mem
Remaining VMs = Minimum (Remaining VMs CPU, Remaining VMs Memory, …)
You can accept any of the available “Change Type”. That is either you use Virtual Machines or Hosts & Datastores. In doing so, you can either change the type of the resource you are modelling with or the way you want to see the future, in this case adding up or modifying VMs in the infrastructure.
You can either use CPU and Memory specific configuration to start the modelling activity or you can use storage based modelling where in you have the option to choose Disk Size, Disk Format (Thin/Thick), Utilization and Link Clone awareness.
If you select to specify a modelling based on your existing VM, you can do so by selecting “Add virtual machines using profiles of existing virtual machines as models“. You have the option to select an existing VM as baseline for sizing data and apply that to the same or to a different datacenter or cluster. When you build a What-If Modelling based on existing virtual machines, you should keep one thing is in mind and that is selecting virtual machines that are similar to the new virtual machines needed in the environment is more useful than estimating the size of the new virtual machines.
Based on what you select for your existing virtual machines, you can specify how often similar virtual machines will be added. The what-if scenario will take these virtual machines into consideration when calculating the overall resource demand. This is a better approach, especially if the average virtual machine approach does not fit the scenario. It takes the basic system configuration like CPU, memory, and disk into consideration, as well as the average utilisation of the virtual machine.
Similarly host based What-If Modeling also offer the opportunity to expand the existing host configuration by adding CPUs or memory, or by adding additional hosts to the environment.
However, Datastore based Modeling has one limitation and that is around the size of the Datastore. The maximum datastore size that you can choose in a model is 2TB. As a work around, you can add additional Datastores of the size necessary to increase the size to the correct levels. Since this is a What-If scenario, using this method simulates the datastore requirements closely enough for planning purposes.
Once you finish Modeling a What-If scenario, details will be saved. Click the circle information icon to see the details of any scenario. From here, you can compare, combine, or remove what-if scenarios.
Please send us your feedback and comments about this #TechTip below, as well as requests for #TechTips on other topics. And yes, follow us on twitter @vcenterops to get the latest on vCenter Operations and Cloud Operations Management topics. For a full list of all blog posts in the vCenter Operations Management Tech Tips series, visit http://blogs.vmware.com/management/tag/tech-tips.
