VMware Novice: 3 Common Traps to Avoid
An Experts Exchange Whitepaper
VMware Novice: 3 Common Traps to Avoid Whether you’re a seasoned IT veteran or just starting out in tech, there’s a lot to know about VMware. Especially if you’re working a full-time job, it’s easy to feel overloaded and overwhelmed. Mistakes are also part of the game, no matter how prepared you think you are. However, if you avoid these three major traps that many novices fall into, you’ll be on a better track towards success:
1. More Doesn’t Mean Better This is one of the more common mistakes that you might make when transitioning into VMware, because you’re applying logic based on previously held assumptions that doesn’t work the same way with VMware technology. Many VMware Junior Admins like yourself are tasked to pump out more performance to meet quarterly goals. In order to achieve that goal, you might think that you simply need to add more power to your systems, whether that translates to video cards, memory, or CPUs for server systems. You might think the same rule applies when addressing power limitations on a virtual server. However, after adding more vCPUs to your system, you’re actually seeing performance drop significantly. How can this be the case? Instead, what you should be doing is scaling out what you do have over more RDS servers and also create
(877) 211-8911 x299
virtual CPUs and memory. By adding more vCPUs, what you’re actually doing is preventing the virtual machines from using the vCPU correctly. By assigning more cores (via more processors) to the virtual machine, the VMware scheduler has to wait until those cores are “free”, thereby pausing these cores and slowing down your overall performance 1.
Unless the software is specifically optimized for more processors, the system will slow itself down.
Unless the software is specifically optimized for more processors, the system will slow itself down. You should instead start by assigning one vCPU to the machine, then slowly increase until there is a performance hit, in which case you should scale back.
2. Relying on Snapshots for Testing Sometimes when you’re trying to avoid making simple mistakes, you can often just end up making different mistakes altogether. As ironic as it may seem, it’s a pretty common occurrence. Let’s say you need to implement a new patch for the version that you have. You’re no doubt thinking of a way to test the patch without messing up your virtual machine. Perhaps in your mind, the quickest way to do that is with the VMware snapshot feature. This allows you to essentially create a picture (or snapshot) of a moment in time from your virtual machine. That way, if something breaks while implementing the patch or new version, you can go back to that moment in time and unfreeze it, as though the problem never happened.
(877) 211-8911 x299
The problem is that these snapshots are not true backups of your system 2. While snapshotting can be used effectively in certain instances, the problem arises when this becomes the go-to procedure for testing. Because the virtual machine will be running slower on the snapshot, the updates will take longer to process. Plus, after it’s done processing, merging the files in order to delete the snapshot will often end up taking hours to complete. This is not an efficient process. Instead, what you should do is take a full clone of your virtual machine to act as a backup before you start running updates. If there happens to be an issue during the process, you can simply destroy the machine, rename the clone, and restart it. If the update is successful, you can simply delete the clone without waiting hours to complete the merge. It’s safer for your system, and faster as well.
Take a full clone of your virtual machine to act as a backup before you start running upda