I’ve written a number of posts on this blog covering how you can perform various forms of capacity analysis with vCenter Operations Manager. This post, about storage capacity analysis, is yet another way that vC Ops can help you with capacity to ensure that you don’t get into trouble. Some of the other posts I’ve created around capacity analysis with vCenter Operations Manager are:
- Introduction to Capacity Planning from the World View
- Capacity Analysis of your vSphere Hosts
- Capacity Analysis of Virtual Machines
When it comes to analyzing any kind of storage, you must always consider that the two aspects to consider are – CAPACITY and THROUGHPUT. These are usually measured in GB or TB (for storage) and I/Os per second (or IOPS) for throughput.
In vCenter Operations Manager, you can monitor storage capacity across all your datastores (we’ll save the throughput monitoring for a separate blog post), do it from multiple angles, and analyze based on today as well as predictions for the future. To gain deeper insight into storage using vC Ops, you can utilize a storage-related management pack (like the management pack for storage devices). Additionally, don’t forget that not only should you monitor datastores but also virtual machine disks, which can have capacity issues as well.
Storage Capacity Analysis From the Operations Tab
One way to look at storage capacity is from the Operations tab. This tab is present on just about any object that you select in the vC Ops Host and Clusters view but it’s contextual. For example, if you select a VM, you’ll see different storage statistics than you would if you select a datastore. For this reason, unless you are focusing on the storage of a particular VM, I suggest you select a datastore and then use the Operations tab to analyze current storage utilization.
Under the Operations tab, you’ll have access to the sub-tabs, Details, Events, and All Metrics. On the Details tab, I suggest looking at what the datastore is “bound by”. For example, this datastore is bound by disk I/O, not storage space (or capacity).
In this case, the VM is bound at 19% by disk I/O and just 2% by disk space.
On the same tab, you’ll find key metrics available on the same sub-tab.
Those key metrics provide you trends for the datastore capacity, provisioned space, capacity as a percentage of workload, and space used by VMs as well as shared space used — all for this datastore.
Finally, still on the same tab, you’ll get a nice overview of the datastore resources, as you see in the following graphic-
These datastore resources tell you about the running virtual machines, running hosts (both on this datastore), total capacity of the datastore, total usage, total VM space usage, total VM space provisioned, total shared space, and then I/O statistics like total throughput and total latency.
If we apply the actual info related to the datastore in the graphic above, we learn that this datastore-
- doesn’t have a lot of VMs relative to the number of hosts
- has a lot of disk space (about 16TB) but very little of that is used (about 380GB)
- provides low latency
What does your datastore capacity look like?
Storage Capacity Analysis From the Planning Tab
The Operations tab provided great information on current storage capacity utilization, but what about the future? The vCenter Operations Manager Planning tab is what will help you to predict the future to ensure that you don’t run out of storage capacity, unexpectedly, and end up with unplanned downtime for your critical applications.
The planning tab offers 3 sub-tabs, Summary, Views, and Events.
The Summary tab helps us predict our long term capacity by providing information like:
Datastore disk space total used – time remaining until it’s depleted as well as the usage trends for the past 2 days, day, current day, next day, week, month, and quarter.
Datastore disk space allocation – provides the same time remaining and capacity utilization predictions but is based on the amount of space that is allocated
Note: if you are confused on used vs allocated, think about thin provisioning where a VM might be allocated 100GB for its virtual disk but it only uses, perhaps, 12GB. In that case, the 100GB would be the “allocation” where the 12GB is the disk space “used”.
You can tweak this view to show this data from the physical vs virtual perspective, how the data is aggregated (average per VM or sum), the perspective (remaining, used, and capacity), as well as the unit to be used (actual value or percentage).
In my opinion, this is the best place that you can go in vC Ops to help you predict disk space utilization for the long term.
Of course, we were looking at that capacity utilization data for a particular datastore. However if you want to go to a higher level, you can see this type of information for ALL of your datastores in one place. Here’s what the same planning tab looks like from the virtual datacenter level-
Here you can see the planning forecast for the entire datastore as well as the cluster level. With this, you can answer questions like “How many more VMs can I add before I’m out of capacity?” and “How long before I fill up my storage capacity?”
If you go into the Views sub-tab, you’ll get additional capacity planning information, as you see in the graphic below-
For example, in the graphic above, the Planning | Views screen tells me that I have 36TB of total capacity and 83% of that is remaining for VMs, which can provide for storage for approximately 62VMs, which, based on my past utilization, is expected to last me for more than a year.
Storage Capacity Analysis From the VM Perspective
Now, let’s jump from the highest level storage capacity analysis that vC Ops provides to the lowest level – the virtual machine disk level. vC Ops provides a surprising amount of VM-level storage capacity analysis. Of course, there is forecasting available on the planning tab but it was the Operations tab where I found the most information.
On the Operations | Details sub-tab, there is some storage information near the top but if you go toward the bottom of the tab and expand out the storage drop down, you’ll get many more details including vDisk level, datastore level, LUN level, and filesystem-level storage analysis (yes I said file system! Note: requires VMware Tools are loaded in the guest VM)
For example, you the filesystem storage details can tell you, for each of your file systems-
- Guest file system free (in GB)
- Guest file system capacity (in GB)
- Guest file system capacity utilization (in %)
Summary
Before reading this article, you may not have realized the deep storage capacity analysis and planning information that vCenter Operations Manager can provide. Stay tuned for more information on analyzing storage throughput with vCenter Operations Manager and for more information on capacity analysis with vCenter Operations manager, see my other posts on the topic:
- Introduction to Capacity Planning from the World View
- Capacity Analysis of your vSphere Hosts
- Capacity Analysis of Virtual Machines
Questions or comments? Please post them below.
Stay tuned for more great real-world, first-hand knowledge about VMware vCenter Operations Manager!