I have a physical server which runs my Virtual Centre but I have heard it said that it is better to run it in a VM.
Has anyone done this? What are the benifits apart from snapshotting the VC?
Waht would the procedure be if I were to move to a VM, (apart from p2ving it)
I think you will get different answers on this. Running vCenter on a VM is supported. We have multiple instances, some are physical and some are virtual. Maybe this will help
You also have the ability to run VM in an HA Cluster - the other option besides p2v would be to do a fresh install creating an ODBC connector pointing to the existing database since all information about your environment is stored in that database and reapp;lying the license to the new license server -
If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
...also, here's a good KB
Hi there
I can just agree, running VC in a VM is supported and works well and you can add also HA and DRS features on it.. Just as my input to this discussion. Everytime I was discussing this question with friends or customers it ended up in almost a "philosophic" question. "should you really virtualize an instance, which is thought to control a virtual infrastructure into the infrastructure itself ?
- I will let you by yourself with this question. Me, often using VC in a VM in demo lab, sometimes also in production. In Projects we often use some older physical servers....
Tom
I have had both virtual vcenters and physical in are environment. I can tell you a vm VC is a pain to trouble shoot when it moves in a DRS environment. Also if you do not have rules setup and VC goes down you get to guess which esx server has your VC vm. I have noticed great performance increases using two physical servers. One for DB and one for VC. This is totally up to the admin that is in charge of the environement. I have 7 blade enclosures fully populated and to mange all of this with a virtual vcenter is a pain. Smaller environments can benefit using a vm but the bigger ones need to stick with physical servers.
I am going to need to disagree with you a little on this one. I have run VC in a VM and Physical. If you are confortable with using the command line finding the VC server should not be a big deal. You could ever write a script and keep it on the ESX hosts to find VC. The point is both ways work and work well. They each have pro's and con's so it would be up to the individual to decide what way will work for them. If you are replicating LUNS to a DR site it is nice to know VC will be there waiting on you.
Steve Beaver
VMware Communities User Moderator
====
Co-Author of "VMware ESX Essentials in the Virtual Data Center"
(ISBN:1420070274) from Auerbach
Come check out my blog: www.theVirtualBlackHole.com
*Virtualization is a journey, not a project.*
I had talked with alot of people at vmworld in Vegas and asked the same question. it seemed to be split down the middle when it came to which was better. We've actually been running it on a vm and have had no issues with performance. HA really takes care of any speed issues that occur.
I'd have to disagree as well. Finding vc is usually not as much of an issue, as a physical vc going down. I run almost 50+ hosts each of two different environments, and my virtual vc performs better than it did on physical. Plus the additional benefits of being able to snapshot before upgrades, and HA/DRS are too much of a plus for me to use it any other way. I've had to revert to snapshots a few times when an upgrade did not go as planned, so that pro in itself was a deciding factor for me, and the other benefits are just icing on the cake.
-KjB
I am currently looking into replicating our SAN for DR so having a working VC instantly appeals to me.
Why / How is it a pain to find the VC if it is a VM?
I am still on the fence with this at the moment, but with the VDI designs I am working on at the moment, I am tending towards a hybrid envorinment with a Physical Master and second level vCenters as Virtual.
It is all about measuring your exposure to risk. VC in a VM instance works and works very well if you do not HA or DRS it you will know where it is on a failure, alternatively have it in a small cluster of two or three hosts then your haystack is quiet small.
If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points
Tom Howarth
VMware Communities User Moderator
Blog: www.planetvm.net
If you have a large environment and DRS moves the VC and a outage occurs you could be in the dark if the VC does not come back online.
You will need to connect to an ESX box with the client to find it and browse the datastores manually.
Not good if you need to respond quickly. But there are ways to prevent this behaviour.
Not a serious issue in my books, but it could blind side you.