Mike D’s Blog

For a long time I’ve told people that if they show me an app then I’ll show them how it can be virtualized. Yes, I’m a little bigoted I guess working for VMware but I do believe there’s not one x86-based workload out there that can’t be virtualized today. To help me prove that point there are 2 great guys that got together to write a book on virtualizing Tier 1 workloads.

This new book written by Charles A. Windom and Hemant Gaidhani just got published last week and it’s available at Amazon. Hemant Gaidhani says: Looks good and worth all the efforts. Click ‘Look Inside’ at Amazon to get sneak preview.

Working with VMware vSphere 4, this book shows you how to virtualize Microsoft applications that require high CPU and high I/O and/or are critical applications for business operations—“Tier 1”applications. With authors who are not only insiders at VMware but who also have developed best practices for multi-tier applications for VMware environments, this book will guide you step-by-step in virtualizing the latest versions of Exchange Server, SQL Server, SharePoint Server, Active Directory, Windows Server, Internet Information Server, and Remote Desktop Services. The authors cover critical topics: reasons why to virtualize the application, considerations to be made when virtualizing the application, setting up a Proof-of-Concept of the application, storage, high availability, and monitoring. Material is organized such that readers can choose which chapters to read, depending on which applications they are considering to virtualize.

Tags: book, VMware

I’ve been catching up on some work and blog reading and ran across an interesting post from Scott Drummonds – one of the most knowledgeable people on performance benchmarking I know. In his most recent post Scott tackles the issue of performance in SQL. This is something that comes up at nearly every customer account I talk to. In nearly every case where I investigate the performance issue it’s due to some sort of misconfiguration. Sometimes the misconfiguration is on the VMware side. More often than not it’s on the SQL side. For some reason people just think SQL is this big hog of resources and the only way to make it run fast and scale is to throw more raw resources at it. Very few people actually take the time to really understand the transactions going to the database, the underlying storage, or any of the other variables that could impact performance of a database transaction.

In his post, Scott goes through some of the simple things you can do to tune VMware for the best database performance possible. For an even more detailed checklist you should read up on the best practice checklist in the VMware Communities. Hopefully between these two nuggets of information you can finally lay to rest the topic of SQL performance.

This post is sponsored by IT Knowledge Exchange. Visit ITKnowledgeExchange.com today to ask your toughest virtualization questions and get answers from your peers.

I got a question from a customer as well as a partner recently about whether or not they should use Perfmon in a Windows VM. I emailed our internal performance team to make sure that my answer of “no – it could give you bad results” was an ok answer. Turns out they’re in the process of writing a KB article on the subject and with their permissions I’ve pasted the draft of that KB article below. NOTE: This applies to ESX 3.5 and prior. Talk to your local VMware SE to get some insight into how things are changing. Unfortunately, I can’t talk about future offerings that haven’t been publicly talked about (press, VMworld, etc) in a blog posting.

Continue reading »

Sounds like something from a technophile’s Christmas list – PowerGadgets. It’s actually a set of tools to create some nice performance monitoring gadgets for the Vista sidebar. I’m mentioning them because yesterday I presented to the VMware R&D group on pains that our customers face when deploying in a Microsoft or Linux shop. This was an internal follow-up to my presentations at VMworld 2008 on the same topics only I tuned this presentation to let our R&D folk know what we could do to help our customers out.

A lot of customers have asked me if there’s a better way to get more performance information out of VirtualCenter or ESX itself. They want some sort of customizable graph they can put on their desktop. Enter this great writeup on how to use the VMware VI Toolkit and PowerGadgets to monitor host CPU and memory utilization. It’s worth a read and possibly even a use or two.

I’m still working with the R&D folk to see what other neat things we could possibly create. In the mean time if you’ve seen or written other nice utilities for monitoring performance then please comment and let the readers know.

A while ago Dell published a benchmark of Hyper-V on some of its servers. The Microsoft people picked up on this and started to say how great Hyper-V was. I really found this interesting because when I read the paper I thought the performance was horrible. I was getting ready to go on a trip to see some customers so stuck the article in Yojimbo to get back to when I had time to write about it. Turns out Gabe beat me to the punch and pretty much wrote down everything I would say although he added a real world customer scenario. The biggest takeaway is the fact that VMware ran VMMark (a virtual benchmark with a good, real mix of workloads) on the same platforms and blew away the Hyper-V results. That test was done before the Hyper-V tests and by completely different people so the results are pretty nice to compare to each other. Anyhow, give Gabe’s write-up a read for the full story.

Unbelievable Hyper-V performance on Dell R900

(Via Gabe’s Virtual World.)

There’s a good write-up of the University of Plymouths Exchange 2007 virtualization project. I just met with 2 customers this past week that both asked about virtualizing Exchange 2007. They both brought up the two biggest questions on Exchange virtualization: what about performance and will Microsoft support it. There’s good answers for both.

Continue reading »