Decisions Decisions Decisions
5/9/2008 -- Comments

With the release of the 2U, 2-socket R805 and the 4U, 4-socket R905 (both AMD based servers) and the already existing 4U, 4-socket R900 (Intel based server) now is a really good time to be looking for a server to run your virtualization farm on. All three are excellent choices and could be the right server. To determine which one depends on lots of things. I'm looking for reasons to choose one over the other (let me know your thoughts). We will be putting together some information to help people decide here on TechCenter and hosting some chat sessions to enable discussion.

Specifically, I ran across this article from The Register that I think has a pretty good take on the R805/R905 Virtualization announcement from a couple of days ago. They point out all of the server options as well as the virtualizaiton services that were a part of the announcement on Tues.

Todd -- Comments




The R805 is Here
5/8/2008 -- Comments

I leave for vacation for a few days and all kinds of cool stuff is announced (although I did make it into the office a couple of times while on the beach). What I think is a really great server for virtualization is now on the website where you can configure and order it. The new PowerEdge r805 is not just a new version of the 2950 or 2970. It was designed with virtualization in mind and has the features to back it up. It has 16 memory DIMM slots, 4 embedded NICs, plus four more PCI slots, and is still only 2U.

I put together a couple of configurations on the website and the price for an r805 system with two top bin quad-core processors and 64 GB of RAM seems like a great deal to me. I guess that I'm still thinking about some of the old PowerEdge 8450 8-way servers that I tested with back in the day with a list price of over 50K, which was a good deal at the time.

Todd - Comments




Hyper-V Quick Migration Part III - Downtime no big deal?
4/28/2008 -- Comments

Jeff from the Microsoft Virtualization team posted part III in his continuing series comparing Quick Migration and VMotion. This entry covers planned outages. His point is that most customers plan for downtime anyway, so having some downtime is fine. He also says that in their conversations with existing VMware customers, they still do their maintenance in downtime windows, even though they have VMotion.

I think this is a fair point, but I also think that customers are wanting to get rid of their downtime windows or shrink them. Quick migration helps to shrink it, and VMotion lets you have only uptime - if you want it.

Todd -- Comments




Microsoft VHD Library
4/22/2008 -- Comments

Microsoft has really demonstrated a strong commitment to virtualization in a variety of ways in the last year. I believe that they are putting forth a very strong effort with Hyper-V which will might be the most compelling feature of Windows Server 2008. In addition to this they have begun to provide preconfigured VMs for development, testing, and evaluation for their server products. Their VHD library has grown to include over 15 products including including SQL Server 2008, Visual Studio 2005, Windows Server 2008, and Vista.

SQLBLog.com has a related post on how to overcome some minor issues that sometimes pop up with these VMs, like the mouse disappearing. As virtualization becomes more mainstream, there will be minor problems like these that pop up from time to time. The time savings with being able to very quickly setup a VM to evaluate new software - is worth it.

Todd -- Comments




Hyper-V Quick Migration Part II
4/18/2008 -- Comments

The second post on the Microsoft Virtualzation Team Blog about Quick Migration and VMotion was added a few days ago. It is mostly a discussion about how VMware HA and Quick Migration both provide a failover solution for UNplanned downtime. I agree with Jeff that both do basically the same thing in the event of an unplanned server outage - the VM is moved to another server and restarted.

The difference that he doesn't talk about is that the rules of the failover are different because of the underlying filesystem that is used in each solution. Microsoft Hyper-V and Quick Migration are using the tried and true Microsoft failover-clustering, which uses an NTFS filesystem on the shared storage. As this is not a cluster file system, the shared storage is actually only visible to one of the servers at a time to prevent corruption. VMware ESX server and VMware HA are working with VMs that are on a VMFS file system which is cluster aware - meaning that multiple ESX servers are able to access the files (or VMs in other words) at the same time.

So following a failure of a Hyper-V server, all VMs on the same disk (or LUN) must be recovered on the same server because the LUN or disk can only be used by one server at a time. In the event of a failure of an ESX server, the VMs can be restarted on any ESX server that has access to the LUN. In order to achieve this flexibility with Hyper-V and Quick Migration it would be necessary to have each VM on it's own LUN. This isn't impossible, but could be more complex to setup and manage.

The end result would be same - the VMs would be restarted on another server. Some of the underlying details can make a difference with flexibility in where those VMs end up.

Todd -- Comments





Virtual NIC Performance
4/17/2008 -- Comments

During our regular TechTuesday webchat, a question came up about virtual NIC performance. The topic this week was Running Enterprise Apps on VMware, and one aspect of that is performance. Hywel, who has a successful SQL Server deployment on ESX, asked - why is he only getting 250 mb/s of throughput? I found a VMware blog post and a whitepaper that indicate it is possible to get much higher. Of course in real-world use through-put will be lower, but I would think that you could top 250 mb/s. I believe that the key to getting the best performance form a virtual NIC in VMware is to install the VMtools and use the VMxnet NIC.

Todd -- Comments




A ScreenShot for Geeks
4/14/2008 -- Comments

I will be the first to admit that I haven't always been a fan of using the biggest baddest system you can find for ESX servers, but some things have changed and I do realize there are cases where really big systems make sense. Which brings me around to the newest system that we ordered for the lab - an R900 with four quad-core top bin processors and 128GB of RAM. This arrived in the lab on Thursday morning and I had ESX on it late that afternoon.

When I saw the system in Virtual Center, it really hit home just how much power you can get in a single system these days. This screen shot is something that only a geek would love (16 CPUs and 128GB of RAM!).



I'm thinking of doing some Exchange 2007 testing.... If I can just find enough disks.

Todd -- Comments




Hyper-V Quick Migration Part 1
4/10/2008 -- Comments

Jeff from the Microsoft Windows Server Virtualization team posted yesterday on their team blog the first in a promised 4 part series on Hyper-V HA and Quick Migration and it's differences with VMware's VMotion or live migration. This first post doesn't really get into the differences, but does a good job of explaining why you need an HA for a virtualizaiton solution.

I am really looking forward the other posts to get the full rundown on quick migration. Jeff makes a reference that this series of posts is in response to "buzz" on the internet. I can only guess that part of that buzz is due to things like a recent video posted by David Marshall on vmblog.com showing Hyper-V Quick Migration in action, and discussions in a few places including the RTFM Education blog, blog.scottlowe.org, and the From the VMware Field Blog.

So it will be good to get Microsoft's side of the story.

Todd - Comments




Is it Cool to Use Virtualization?
4/9/2008 -- Comments

These days, it seems like everybody is at least looking at using VMware or Xen or some other virtualization solution. It's the cool thing to do. Clearly.

But what about when you run something like Exchange or SQL Server or Oracle? It seems that almost everybody is cool with virtualization, until you start talking about stuff like this. Then there is a pause. Some decide to push forward, others decide to wait, still others may declare that they would never do it!

There are lots of issues to consider including support, performance, backup, disaster recovery, and even non-technical things like organizational structure and management support. So the answer for everybody isn't the same. The next TechTuesday chat is going to have this as it's discussion point - Running enterprise apps virtualized - to let everybody learn from each other.

Todd - Comments





SQL Server on VMware Server
4/4/2008 -- Comments

I recently ran across this blog entry on Scott's Blog (no relation to Scott Hanson on my team) about a migration of a SQL Server DB to a VM running VMware Server. I have spent only limited time working the the VMware Server (formerly GSX) product and instead have concentrated on ESX. This entry has some interesting info about their experiences with getting SQL Server production DB performing at acceptable levels including a nice "lessons learned" section at the end. For those thinking about going the VMware Server route, then this is a great read.

Todd -- Comments





Running SQL Server in a VM
4/2/2008 -- Comments

I published a new white paper this week on running SQL Server in a VM. The paper focuses on the performance of a large (well 100 GB) version of the DVD Store when running on 64-bit SQL Server 2005. I started this testing just after ESX Server 3.5 was finished and made available, and I believe that it is the first paper with performance numbers of SQL on ESX 3.5 (Please correct me if there are some!). In addition to the paper we had a web chat on Tues on the topic. And I have setup a wiki page to collect SQL on VMware resources and serve as a central place to host threads.

A couple of interesting notes, well at least interesting in a kind of really geeky way:

The first performance study on VMware that Dell published was a paper by Dave Jaffe and I on VMotion performance. We used a 1 GB version of the DVD Store database on SQL Server 2000 as the test VM. So this test really is an update in some ways to that paper. Same basic conclusion regarding VMotion with SQL Server - it is very hard to measure or detect the impact to DB performance of a VMotion.

This study was similar in many ways to the testing that Kong Yang and I did late last year on Exchange 2007. Key difference is that SQL Server is much easier to test! Exchange and LoadGen have extremely time consuming requirements to setup and initialize testing, but the DVD Store on SQL Server can be created and loaded in a much shorter time, and includes cleanup or resetting scripts that are also very fast.

Todd -- Comments




PERC6 Performance Analysis Report
3/28/2008 -- Comments

Our System Performance Analysis lab recently completed a series of tests to evaluate the performance of the newest PowerEdge Expandable RAID Controller (PERC) and published the results in a performance analysis report. The tests with the PERC6 were very through with many RAID types tested - 0, 1, 5, 6, and 10 - on 12 different workload type profiles. Results provide guidance for what PERC6 settings will provide optimal performance for a given workload type.

Additional cool stuff in the report: A good discussion about all of the different RAID types with definitions, and an analysis of the results which recommend when to use which RAID type. The guys that do these tests are storage geeks - and it shows in the report!

Todd -- Comments




Using Simple Password on WS08 Domain Controller
3/24/2008 -- Comments

In the process of getting my new Windows Server 2008 Domain Controller VM setup, I needed to set the password for the new domain's Administrator account to match our lab standard password. This password doesn't meet the default complexity requirements that the Domain Policy establishes. So I had to use Group Policy Management (located in Administrative Tools) to edit the Default Domain Policy. In Computer Configuration > Security Settings > Account Policies > Password Policy, there are several settings that control what passwords are allowed for how long. I disabled the complexity requirements (must have a capital letter, lowercase letter, number, and not be based on a star trek character type of requirements).

I then saved the changes and verified that they had taken effect by viewing the domain policy. I was already logged on as the domain Administrator, so I pressed CTRL-ALT-DEL and selected Change a password.... I typed in the old password (that I had made suitably complex to be able to complete installation) and the new simple password. I received an error message "Unable to update the password. The value provided for the new password does not meet the length, complexity, or history requirements of the domain."

After checking again that the domain policy was in fact changed to allow simple (or non-complex) passwords, I did what any good Windows admin would do - I rebooted. Following reboot it still did not allow me to change passwords. I created a new account - Admin - that way a copy of the original Administrator account and was able to assign it the simple password without a problem. I then logged on as this new Admin user and used Active Directory Users and Computers to reset the Administrator account to the new simple password - with no problem.

So I wasn't able to change it from the CTRL-ALT-DEL (or windows security) Change a password... screen, but I was able to change it by resetting it from another account. I wasn't able to find any other hits via google on this problem and resolution, even after I had figured it out. I'm not sure what this is the way it works, but I did find a way around.

Todd -- Comments




Opps Leads to PSOD and New WS08 Domain Controller VM
3/20/2008 -- Comments

I have a confession to make. I was trying to rush through a setup of VCB server on our SAN late last week and ended up reformatting the wrong LUN. Now this was all in our lab so no matter what gets destroyed - it's not the end of the world. But of all the LUNs, this was probably the one that I really didn't want to get reformatted. It had the Domain Controller for the older of our two Windows Server 2003 Domains in the lab. This DC was also the only DNS for that portion of the lab - so DNS was out which was the really annoying part.

It took me about a half a day of investigation to figure out exactly what had happened because that older domain isn't used as much and a weekend passed before anybody complained. At first I thought that the VMFS partition had gotten corrupted, which turned out to be kinda true. The problem was it was me that corrupted it by formatting it with an NTFS filesystem.

After mourning for a few minutes, I decided that this would be a good opportunity to install my first Windows Server 2008 domain. First I straightened out the LUNs and formatted the correct LUN as NTFS and unassigned the initial LUN away from the windows system. I then used the Add Storage option in Virtual Center for the ESX server that I was using to get the LUN reformatted as VMFS. For some strange reason I was then getting an error when I tried to create my new WS08 DC VM on that LUN. So I tried to create it from another ESX server in the same server farm. I had to do a rescan of the storage adapters, but then the storage showed up and everything seemed to work OK.

I used the WS08 RTM ISO to install the new VM, but got an error on reboot of the VM that the installation was corrupt. I did some investigating and found that several ESX systems thought that they still had VMs running from the original LUN before it was formatted NTFS and then reformatted VMFS. This ended up causing a PSOD (Purple Screen of Death - very cool if you haven't see one yet) on one ESX server when I tried to access one of these VMs. It apparently also caused some corruption in the new VM that I was building.

So I went back and did storage adapter rescans on all ESX servers in the farm and made sure that all VMs that were from the LUN that I deleted were removed from inventory. After this my install of the new DC went great.

I have only had two PSODs while working with ESX, this one was caused by me doing something that you really really shouldn't do with your storage. I learned (or maybe learned again) that you should always take your time when you are going to do something like format and make sure you have the right disk!

Todd -- Comments




VMware Blog Entry on Exchange
3/7/2008 -- Comments

DellTechCenter's resources on Exchange on VMware were included in a post on VMware's blog. The entry is about the recent testing VMware did to get 16,000 Exchange 2007 users on a single server by using ESX Server. The entry then goes on to cite other Exchange on VMware resources and includes our very own TechCenter resources:

Exchange and VMware at the Dell Enterprise Technology Center wiki

• Virtualizing Exchange Server 2007 on ESX 3 - Whitepaper
• VMotion and HA with Exchange 2007 on ESX 3 - Whitepaper
• Virtualizing Exchange 2007: The Final Frontier - Presentation
• Exchange Server 2003 Performance on ESX 3 - Whitepaper

They have lots of other great info on Exchange on VMware included in the post. It is recommended reading.

Todd -- Comments




VMotion of Exchange Mailboxes on ESX Server
3/5/2008 -- Comments

I saw this blog entry on savagenomands that wondered about VMotion with large Exchange VMs:

"Even with all that horse power, they didn’t talk about network traffic running all exchange 2007 roles on the same box. I wonder how long one of these boxes takes to vmotion. Also I’d like to know the details on how they setup the luns and if they used vmdk files "

I actually did some similar testing with 2000 user Exchange VMs and measured the time for VMotion to complete as well as its impact on the users. The answer - about 6 and half minutes to VMotion a 2000 user VM while under stress (At least with the setup that I had. Your mileage will vary depending on your exact configuration). See the whitepaper on the VMware and Exchange page to get the complete details.

Todd -- Comments




Dell EqualLogic Blog
2/29/2008 -- Comments

I had the chance yesterday to meet Marc Farley who is a new to Dell via EqualLogic. He is running a blog on the Equallogic site. Marc participated in the EqualLogic webchat we did on DellTechCenter a couple of weeks ago as a subject matter expert. So I had kinda met him before, but this was the first in-person meeting.

We had a great discussion about the things they did at EqualLogic to connect with their customers and we talked about some ideas for continuing that now that EqualLogic is part of Dell. We are currently in the process of getting some of the new EqualLogic arrays into the TechCenter lab. So let us know what kinds of things you would like to see / hear about.

Todd -- Comments




Firewall VMs and Virtual Switch Security
2/27/2008 -- Comments

On the Tech Tues chat yesterday a question came up about setting up a firewall VM by connecting one of the NIC ports on the ESX server to external Internet. The question was - Is this a good idea? Is it secure? What is the best practice?

I did a little research this morning and found a white paper from VMware that mentions the exact scenario of having one VM connected externally and acting as a firewall. On page six of VMware Networking Concepts (located at http://www.vmware.com/files/pdf/virtual_networking_concepts.pdf ) there is a diagram and discussion about it.

There is another VMware doc titled ESX 3 Best Practices Guide that I found attached to a VMTN discussion ( http://communities.vmware.com/message/861839#861839) that covers networking best practices. It also has the same recommendation that using virtual switches is very secure and allows for things like firewalls to be setup in a VM.

Todd -- Comments





New Storage VMotion Demo
2/21/2008 -- Comments

I had a chance to play around with the new Storage VMotion feature of VMware ESX Server 3.5 / Virtual Center 2.5 and found it to be easy - once you get the command line tools in place. You have to have the Virtual Infrastructure Remote CLI package in place to be able to execute the svmotion command.

I was able to move the SQL Server VM from one LUN to another while under stress using Storage VMotion. I think that it's kinda cool to see in action - so I captured it in a quick video.

Yes - That's my voice on the demo - so be kind with your comments!

Todd -- Comments




Virtualizing Big Databases
2/20/2008 -- Comments

I've been doing quite a bit of work over the last few years to provide some data as to how things like SQL Server databases and Exchange mailbox servers perform inside a VM. All in all the story has turned out to be pretty good. Sure, running in a VM does have a certain performance overhead or penalty, but when most servers are running at 20% utilization or lower - this turns out to not be a problem.

Specifically I have been running some SQL Server 2005 tests, which I am wrapping up now, in the last few weeks. On a conference call today where we were discussing the results a question came up that I thought was pretty good -

What "size" of SQL Server database do most people run?

Of course this is an impossible question to answer, but of course I came up with an answer anyway! In terms of SQL Server databases that are running in VMs - I think that so far most people have only migrated "smallish" databases. Ones that don't have extremely high transaction rates. I do know of a few cases where more performance intensive databases have been virtualized, but these are the exception today. I think that in the future more of these high performance DBs will also be virtualized.

I think that many IT shops are looking at virtualization and think that it would be really cool to move some of their "bigger" stuff into VMs. To take advantage of a better way to manage things, but they are waiting to see how everything plays out before they begin to virtualize big apps like SQL Server, Exchange, Oracle, etc.

Todd -- Comments





Performance Testing with Hyper-V RC1
1/23/2008 -- Comments

I wrapped up a quick round of performance testing with the Release Candidate 1 of Hyper-V (Included with Windows Server 2008 RC1) using the same test that I used with RC0. The initial test with RC0 was just to measure how well Hyper-V distributed the system resources among 8 virtual machines that were under load. In the case of RC0, and now in RC1, it appears that Hyper-V does a pretty good job of evenly distributing these system resources. The test showed that the variation in performance between the 8 VMs was small.

By using the same test (same server, storage, and test application) it is now also possible to compare the performance of RC0 vs RC1. The performance guy from the Hyper-V team at Microsoft claims on their performance blog page that performance is improved with RC1, which were able to confirm with an increase of 3% in Orders Per Minute and a 5% decrease in average response time. These percentages are not large, but they are improvements!

Todd -- Comments




Perfmon with Hyper-V
1/15/2008 -- Comments

While testing and evaluating the new beta of Hyper-V that is included with Windows Server 2008 I have run across some interesting performance monitor stuff.

First a little background - I have been testing and evaluating the performance of VMs on VMware's ESX server since early 2004 with the first whitepaper I'm aware of that took an detailed look at the performance of VMotion. So I have been using esxtop ever since to monitor the performance of ESX servers.

Windows Server 2008 RC1 has a new Perfmon that is much improved over previous versions, so I was really looking forward to using it for my Hyper-V testing. The first testing that I did with W2K8 RC0 showed that Hyper-V (called Windows Server Virtualization at the time) was able to evenly distribute limited resources across 8 VMs all under load. When using perfmon to monitor the CPU during these tests - it at first looked like the server wasn't loaded at all. I was looking at the standard %Processor Time for the entire server. I knew that the load I was running should be pushing the server well over %50 utilization, so I started digging into this new perfmon to see what was going on.

I found a set of counter objects that were labeled as hypervisor. Here there were Hypervisor Virtual Processor, Hypervisor Root Virtual Processor, Hypervisor Root Partition, Hypervisor Partition, and Hypervisor Logical Processor. It seemed that surely one of these would give me the overall CPU utilization that I was looking for, but it wasn't really clear which one was the counter to look at. As this is still beta code, there was no documentation for these counters to clearly state what each one meant. I decided in the end that the CPU utilization wasn't really that important to report as part of my results, and I just decided to look at it again when the next beta was released.

It seemed like a big issue to me because it would be possible for admins to look at %Processor Time and see a low utilization number while the system was actually very heavily loaded with several very active virtual machines.

With the Hyper-V beta that is included with W2K8 RC1 (released in Dec 07) I am currently doing a new round of testing to see if there is any change in performance. I'm doing the same test with the DVD Store database, but this time the perfmon results are very different. The %Processor Time for the server now seems to reflect the activity of the virtual machines (or partitions as they are referred to in perfmon).

I'm still working on finalizing the results of the RC1 testing - I'll have an entry here when they are ready.

Todd - Comments




Disk Management in Windows Server 2008 RC1
1/8/2008 -- Comments

Immediately after finishing the post on performance testing of Windows Server Virtualizaiton testing, Microsoft released W2K8 release candidate 1 with a updated version of their virtualization layer. The virtualizaiton role is now referred to as Hyper-V.

I've set out to do a new round of testing with Hyper-V and W2K8 RC1. However, I ran into a problem getting connected to the iSCSI LUN on the MD3000i. I was unable to get the Disk Management tool to load from within Server Manager. At first I thought that it was an issue with how I setup the iSCSI initiator on WK8, but some google searches revealed that this was a more generic issue with the Disk Management tool. And infoworld reviewer Paul Venezia ran into the exact same problem as me during his review.

I was able to workaround this by using the DISKPART command line utility to mount the drive that resided on the MD3000i. I was using the same LUN as with my previous testing, so I just needed to use the DISK ONLINE command to get my drive up and running, but there are other commands such as CREATE and FORMAT that can be used to get a brand-new LUN partitioned and ready for use.

I was also able to boot the VMs from the previous testing, proving that they are portable. I will most likely recreate new VMs however, to evaluate the create VM wizards in this new version.

Todd - Comments




Windows Server Virtualization Preview Performance Testing
12/13/2007 -- Comments

I recently took the Windows Server Virtualization preview (included with Windows Server 2008 release candidate 0) for a test drive. The intent was to get it installed, create some VMs, put them under load, and see what happens. As this is still pre-release it isn't fair to compare the performance in any absolute terms with either physical servers. But I didn think that it would be fine to put it under some stress and see how the new hypervisor handles the load.

The net results are posted on the wiki in the Microsoft Virtualization section. To sum it up shortly - the system handled the stress very well.

As new releases of the virtualization technolgy to be included with W2K8 come out, we will evaluate them and post what results we can. Let me know what types of things you would like to see.

Todd -- Comments




VMotion and VMware HA with Exchange 2007 VMs
11/29/2007 -- Comments

A second whitepaper from the Exchange 2007 on VMware testing that Kong and I presented at VMWorld has been published on the wiki. It covers the performance of 500, 1000, and 2000 user Exchange VMs under load while using VMotion and VMware HA. The complete results are in the paper, but a quick summary is that all VMs performed without error in all the VMotion and VMware HA tests.

At our presentation at VMworld we talked about how we saw some unexplained behavior with the large VM with both VMotion and VMware HA. We found that by lowering the amount of RAM assigned to the VM, so that one GB was available for ESX to use exclusively, that the performance of the large VM with 2000 users was much better.

A landing page for Exchange on VMware has been created on the wiki that has all of the whitepapers and the VMworld presentation.

Todd -- comments





Whitepaper on Exchange 2007 and ESX 3 Performance
11/7/2007 -- Comments

The whitepaper based on the presentation from VMworld has been published in the most recent edition of Dell PowerSolutions magazine. The paper has more details on the results from our testing with Exchange 2007 VMs. It focuses on the performance testing and characterization work that we did. The basic conclusions were that performance was acceptable for the 500, 1000, and 2000 user workloads that we tested. Additionally, in order to optimize performance you should only use the number of virtual processors that are needed.

Check out the full details in the paper and comment here if you have any questions or want to share your own experiences with Exchange in VMs.

Todd -- Comments




VMWorld 2007 - V3i and Exchange 2007

10/24/2007 -- Comments

I recently attended and presented at VMworld 2007 in San Francisco. It was an absolutely awesome conference with lots of great sessions, tons of attendees, and also fun afterhours parties. One of the big announcements of the show came during Diane Greene's keynote on the first full day of the conference where the embedded version of ESX 3 – V3i – was shown running on the unreleased / unannounced Dell r805 server. A slimmed down version of the VMware ESX server hypervisor is booted from an internal flash card. On stage Diane and Mark Jarvis from Dell showed how quick this combination boots and how easy it was to create and run VMs. The entire keynote has been made available on the VMworld website. (The demo starts at about 10:45 and ends at 17:15 )

Look for more information about the r805 soon.

I presented a super session on virtualizing Exchange 2007 with Kong Yang. We had a great audience and almost everybody stayed until the end where we had lots of great questions. Key findings that we presented were that good performance was attainable with Exchange 2007 in VMs at all sizes that we tested (500, 1000, and 2000 users). We also found that performance of Vmotion and VMware HA was better with the 500 and 1000 user VMs than with the 2000 user VMs. The complete presentation is available on the wiki on the VMware and Exchange page.

Since returning from the conference I spent some time trying to understand the performance issues that we found with the large 2000 user VM in the Vmotion and HA tests. It turns out that we were not leaving enough memory available for the ESX host, which resulted in swapping. Kong and I have written a white paper with the details and it will be posted soon.


Todd -- Comments




Welcome!
10/18/2007 -- Comments

This new blog on Dell TechCenter will focus on the leading edge of virtualization for the enterprise. Server virtualization over the past several years has been getting more and more interest. There seems to be new announcements coming out all the time about new products, offerings, technologies, partnerships, and agreements.

There is a lot going on and there are lots of questions. This blog will focus on the technical aspects of virtualization on Dell server and storage products. We are going to provide information about virtualization that comes directly from the Dell labs AND we are inviting the technical community to join us.

Ask questions, provide answers, and get a better understanding of the virtualization landscape.

Todd -- Comments