Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

first I'd like to say I'm relatively new to this site. Wow.. This is AWESOME!

My very first question pertains to virtualization. We are a trading firm that relies heavily on low-latency networking and high-performance computing, mainly running SQL. However, we do have several candidate servers that we're seriously considering virtualizating to VMware ESX.

I have a background with VMware, but not so much the P2V side of things. Where I'm having difficulty is initiating the process of baselining the systems that we'd like to virtualize. We are using SolarWinds Orion product to monitor out systems performance. I assume this will be a vital tool in throughout the process.

I'm just looking for any others experience with P2V, and the methods used to baseline the servers.

Many thanks!


share|improve this question

Bit of a sweeping statement I know but in general I'd never virtualise an even remotely-busy database machine in production, for test/dev sure but they'd be literally the last class of machines I'd virtualised - tread carefully.

share|improve this answer
Agreed. IMHO a high performance database should not be virtualized (yet?). There's always a (relatively) small-ish performance penalty with any virtualization product. Until the performance practically matches the hardware w/o ESX a "trading firm that relies heavily on low-latency networking and high-performance computing" might want to virtualize other servers (AD,File,Print,etc.etc.) before the database servers/cluster. – osij2is Jul 21 '09 at 21:52

Every VM conference I've been to has the VMware techies claiming that there is no type of server that can't be virtualised. They claim large Oracle VM installations that handle 3 times the amount of daily Visa credit card processing (this is with VShpere 4.0).

But they live in an Ivory Tower. For us in the real world I would avoid virtualising database servers. Otherwise the DBAs will start accusing you of "slowing down their server" every time something happens (whether it is VM fault or not). It's not worth the hassle.

share|improve this answer

If you currently do load testing you're in luck.

Build a virtual QA and compare!

I think you're doing the right thing since SQL virtualization is inevitable.

SQL DBA's will not love you though. You throw in variables like host VM load, SAN bandwidth, Microsoft support issues and network security. It's getting complicated.

share|improve this answer

If you talk to your VMWare Solution Provider they should be able to point you towards a utility you can install that will help you benchmark the important items.

share|improve this answer

Assuming the servers that you are considering virtualizing are running Windows, you'll want to try out the Microsoft Assessment and Planning Toolkit. It's free, does an agentless (and comprehensive) inventory of everything in your environment, and amongst other things, will give you a detailed assessment and reporting of server utilization gathered using the Performance Metrics Wizard, and will provide recommendations for server consolidation and virtual machine placement.

Also cool is the Power Savings Assessment which creates a proposal to identify server and client machines running in your environment and understand the power management capabilities available.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.