VMTurbo enhances our ability to manage our virtual infrastructure
July 22, 2016
VMTurbo enhances our ability to manage our virtual infrastructure
Score 9 out of 10
Cloud Edition,Enterprise Edition
Overall Satisfaction with VMTurbo Operations Manager
We use VMTurbo to monitor and manage our VMware environment. Currently it is used by our sysops group exclusively, but we are working on ways to use the reporting capability to keep our executives apprised of the current status of our environment. The main reason we started looking at VMTurbo was because we did not have a good picture of what exactly was using all our resources in VMware and did not have a good way of seeing how the usage of resources by certain VMs was affecting the overall environment.
- One of the most useful things we have found with VMTurbo is that we are able to see if we have over/under provisioned resources. Because VMTurbo keeps track of both average and peak usage, we are able to confidently assess when a VM needs more resources, or if we can take away resources and use them elsewhere because the VM was not properly sized in the beginning.
- The automatic resource management is so much better than vCenter. When we set a cluster to have its CPU/Memory managed by VMTurbo, we see an immediate improvement in how the VMs are spread across ESXi hosts. This improves performance since VMTurbo is taking care of any hot spots and dynamically adjusts resources so that things are running as efficiently as possible.
- Being able to get an overall view of what is consuming resources in VMware has been extremely helpful. We have seen instances so far of VMTurbo saying that we have plenty of resources in a cluster and we can remove a host, and also of cases where we the only thing you can do to remove resource contention is to add hosts to a different cluster. Being able to have statistical data to back up our decisions (which were normally based on hunches/experience) has made it easier to justify additional physical resources when needed.
- The default retention for data in VMTurbo is 12 months. We have cycles in our environment that are quarterly and annually. We would love to be able to have multiple years of information easily available to us since our management is always wanting 2-3 years worth of peaks to assist in making hardware decisions. We have been told that we could expand the data retention somewhat through support, but that it may have impact on performance of VMTurbo. An easy way to expand the reporting capability to cover 3-5 years of data would be helpful.
- The terminology and workflow of VMTurbo is a little overwhelming at first. It really is a very unique way of looking at things in your virtualized environment. This is not necessarily an area where they need to improve, I believe that once you understand it, it becomes very powerful in a way it could not be using just the old way of looking at things. My suggestion however is to ensure that you get and take advantage of the training/onboarding they provide so that you understand the program. Without that, many of the features will never be used because they are new to a standard virtualization admins skill set. However, they are always willing to spend time with you to make sure you understand how to do things...they are very helpful.
- VMTurbo has had a big impact on our ability as an Infrastructure Support Team to identify areas where we are wasting money. One example is that we had 14 specific app servers that each had 64GB of memory assigned to them. We suspected that they were over provisioned, but the reasons we were giving for our suspicions were not accepted very well by the business. They wanted to ensure that their app would perform well, and they felt that they needed those resources to do it. After running VMTurbo for a few months, we were able to show them that we had way too much memory provisioned for those servers and we were able to consolidate down to 7 servers with 64GB of memory per. The next step is to show them that each of the servers can now go down to 32GB of memory, but we go 1 step at a time. We are finding many other instances where VMs are over provisioned and this is going to allow us to actually purchase less hardware this next year.
- We have had multiple instances where we have used VMTurbo as a troubleshooting tool in order to find issues in our environment. For example, we were having some storage performance issues on a certain datastore and we were able to go into VMTurbo and see which VM on that datastore was using those resources the most. It turned out to be a VM that should never have been using that much, and after working with the app admin we found that a process was freaking out and not working the way it was supposed to. They fixed that and the usage went down considerably. Without VMTurbo, it would have been much more difficult to find the particular VM that was impacting the datastore.
The automatic placement of VMs is a really cool feature of VMTurbo. I have found that vCenter does a really poor job of deciding through DRS what VMs should be where. We would frequently find a cluster where 2/3 of the ESXi hosts were using about 80% of memory, and the other 1/3 were only at 15%. In addition some of the VMs on the busy servers were ballooning and swapping memory, while there were plenty of resources available elsewhere in the cluster. As soon as we turn on VMTurbo automatic placement on a cluster, we see results. Usually we find that by the next day, all the VMs in that cluster are evenly distributed (based on utilization) across the ESXi hosts. We have not been able to automate all of our clusters yet, but the business is starting to see the benefit of that and allow us to automate things where they were not willing to before because of concern over Change Management and things like that.
We have used vCenter and vFoglight for management and also evaluated Solar Winds and Cloud Physics. Each of those systems had some pros and cons, but overall we found that VMTurbo had the things that we wanted and also things that we didn't even know were possible. We had been using the Foglight tool mostly, and in general, I could configure it to do many of the things that VMTurbo does, but it took quite a bit of time to do it. We don't want to have a person whose only job is to manage the software to manage our virtualization environment and who is the only one who even understands how the thing works. We just found that VMTurbo had the right combination of integration with VMware and visibility into the environment.
If you have a small, simple virtualized environment, VMTurbo may not give you a whole lot of added value. If, however, you have any performance or resource management issues in your environment, no matter how small it is, you would benefit from having VMTurbo. Our virtualization infrastructure, like many, is a combination of different speeds of storage, VMs of different sizes and performance needs, and it is always growing. In a scenario like this, VMTurbo has been able to give us visibility into what is using resources, what is needing more resources, and how all these resources can be arranged so that we are using the physical resources in the most efficient way possible. Because it is looking at the whole environment, it also gives us the leverage and statistics we need when it is time to request new or different hardware...sometimes more hardware IS the answer, and when that is the case, VMTurbo helps you to quantify that.