I had the poll automatically complete last friday and here are the results: Of 110 votes cast by my readers, 57 of you said you run vCenter in a VM, 50 run it on Physical, and 3 don’t run vCenter at all.
Great information and thank you all that responded to the poll.
How do you run vCenter (Virtual Center)?
Total Voters: 110
Loading ...Posted a new poll this past weekend. I see some have already started responding. I know a few blogs have run this one in the past but I’m curious with my SMB focus to see if the results vary. Please take a quick sec and post a vote. I’m closing the poll on March 7th (an extra week because I started it late).
Just as a FYI, I run mine in a VM.
I’m a firm believer that better decisions are made with better information. When you open up the Virtual Infrastructure Client and select Administration/VirtualCenter Management server configuration. From there select statistics in the left pane. You will see a right pane that looks something like this:
By default, Virtual Center has set the statistics level to 1. The levels can range from 1 to 4. Level 4 maintains the most stats on your environment and is typically used for debugging. Continue reading »
I get asked this question at least once a week. Should I run Virtual Center as a physical machine or as a Virtual Machine? The true answer is: either. I personally think it’s better to run as a virtual but it also depends on how comfortable you are with your Virtual Infrastructure.
You can run VC as a physical machine. Most of my larger customers run it that way. Their idea is that, "I don’t want to run my Infrastructure Manager in the Infrastructure it’s managing." I get that. It the previous versions, I might agree.
If I were going to build an virtual infrastructure today, I would run it as virtual. Here’s why: all the benefits of virtualization. For instance, I can take a snapshot of my VC server for Business Continuity and/or testing. I can have DRS balance my VC server on the infrastructure nodes that are most available. Most importantly, and one some customers don’t realize, is that High Availability will fully function. Even for the VC Virtual Machine itself. If the physical host where the VC server is running goes down unexpectedly, HA agents running on ESX itself will power on the VC VM elsewhere in the cluster. VC is only required to be running to setup and configure HA. Once configured, the agents running in ESX monitor and restart the VM’s in the case of a host failure. Of course, during the restart of the Virtual Center VM, DRS and VMotion will be unavailable during the boot cycle of the VC server. In my opinion, a small price to pay compared to the physical world. If VC were physical, how long would DRS and VMotion be down for? 30, 60 minutes while someone dialed in or drove in at 3am? In the virtual world, I could envision an outage of DRS and VMotion for 3-5 minutes while the VC server booted on another host.
Running VC in a virtual is not without it’s challenges. You have to manage the VC VM a little differently. To make changes, or move it, or power it down for any reason, you have to manage it from the physical host it is running on. So after powering the VC VM down, you have to connect the Virtual Infrastructure Client (VIC) to the host the VC VM was running on. From there you can make modifications to the VC VM, change the ram, power it on, etc. These are all things that are learned over time managing a virtual Infrastructure. If you’re not ready for that, you could start VirtualCenter as a physical server as as you become more virtual savvy, you can run the standalone version of VMware Converter to convert the VC server from physical to virtual. That’s always a possibility as well.
As always I recommend reading some documentation (RTFM) to fully understand this decision. VMware docs for virtualizing VC and all of it’s benefits and caveats can be found here .