I began my current career path as an IT consultant back in 2005. If you recall, this was just about the same time as VMware started raising eyebrows. The release of ESX 3.0 came the following year, 2006. And thusly, the server virtualization revolution went into full effect.
We had a lot of challenges those first few years. Some of those are still around today. One of those challenges was, and still is, capacity planning. Before VMware released their Capacity Planner we were stuck using tools like IBM’s CDAT (Consolidation Data Analysis Tool) that did nothing more than just pull performance data via WMI and dump it into a CSV file. I still have nightmares of having to troubleshoot missing Windows perfmon counters, ADMIN$ shares that had mysteriously been disabled, and having to deal with all of those machines that weren’t on a client’s AD domain for no apparent reason. Those days are long gone and we have MUCH better tools available to us for such sizing efforts. But now we face new challenges.
Clients are now on their 3rd, 4th, 5th, or even later generation server virtualization platform. When clients perform these platform refreshes, they typically (based off what I have seen) only refresh the underlying server platform and install the latest version of their hypervisor of choice. Few, if any, look at the performance of their existing VMs and make any determination as to whether or not those VMs need more or fewer resources to meet their application performance requirements. This is a difficult task to assess. You can’t just approach this challenge the old fashion way and look at averages of perfmon counters in order to size appropriately.
Another challenge that has entered the marketplace is a greater use of storage array-based snapshots as a means to augment a client’s data protection strategy and sizing properly for such use. As many of you know I’m a huge proponent of NetApp storage. In particular, I’m a big fan of using the more advanced snapshot features on Data ONTAP-based FAS arrays in order to provide for fast recovery. I’ve seen many times recently where clients have recovered from such things like ransomware only because they had hourly snapshots in place for their NetApp-hosted file shares. Same goes for failed patch attempts on VMs hosted on NetApp. Snapshots offer so many benefits. Don’t get me wrong though, I still believe you need to have a second copy of your data elsewhere and another copy offsite to meet the industry standard 3-2-1 backup rule. Snapshots are only one part of a greater data protection strategy. But let’s not digress…
When I first heard about Veeam ONE, I viewed it as most people: as merely a monitoring tool for value-add sale on top of what Veeam was offering from a data protection solution. My mind was changed when I saw one of Veeam’s senior Solution Architects, Rich Brambley (@rbrambley), use this as a planning tool while helping us with a larger data center transformation for a client with a few dozen data centers. Veeam ONE comes with a number of canned reports that do an incredible job of showing resource consumption on a per VM-basis. Moreover, it gives recommendations as to what resources should be assigned to a given VM. Both the “Oversized VMs” and “Undersized VMs” reports are incredibly helpful in this regard. These reports allow an architect to see exactly what additional, or fewer, resources would be required when building out the configurations for new hypervisor hosts.
Veeam ONE also has a number of traditional Capacity Planning reports as well. This includes reports for “Host Failure Monitoring” – determining what resource consumption looks like after a failure of one or more hypervisor hosts – and the “How Many More VMs Can Be Provisioned” report for a given cluster. With these reports in-hand, sizing a hypervisor host upgrade becomes incredibly easy.
As previously mentioned, another challenge that I have encountered is how much capacity to account for in a NetApp FAS design when a client intends to use NetApp snapshots. In the past, our only real resource for such sizing was to look at a client’s backup reports and see how much data was backed up during each night’s incremental backup job. Even then these numbers were hardly at all accurate. Veeam ONE has taken a lot of the guess work out of this effort. Bar none, one of my favorite reports from Veeam ONE is the “VM Change Rate Estimation” report. The original intent of this report is to assist with sizing out Backup Repository storage for Veeam Backup & Recovery. However, the same report can be used to determine the amount of storage that could be consumed in snapshots over a given period of time.
As you can see, Veeam ONE is a very powerful tool. It does a great job as monitoring and alerting tool as well. I don’t want to undersell that fact. However, it has been an instrumental tool in assisting me with upgrading clients’ existing servers and storage.
Please Note: I do not work for Veeam, get any funding, nor am I compensated in any way, shape, or form for the writing of this article. This is by my own accord and will hopefully be helpful to others in my type of role.