Blog Posts

During our weekly tuesday chat, I told ye110wbeard that I would send him a TechCenter shirt if he created a theme song for Hyper-V. I felt secure that I would not have to actually payout. Unbelievably, he had a Hyper-V theme song and video and lyrics up by late Wednesday. I think the song is kinda catchy and the lyrics are pretty good. Check it out when you get a chance. And I am working on sending out not only a nice TechCenter shirt, but also a nice coffee mug tumbler thing as well because he went above the original challenge and added video!

Todd

During last week's chat the we discussed ESXi licensing and features. In the features portion of the discussion flakrat asked if it was possible to setup two ESXi servers with shared storage to be able to do a manual fail-over of a VM between the two servers. I was able to create this setup and confirm that it works. Read on for details.

The setup that I used to test was two PowerEdge R805 servers with ESXi Update 2 refresh installed on the hard drive. This was the currently available "free" ESXi installable from VMware's web site. To install I used the Dell Remote Access Card (DRAC) virtual media capability to boot from the ESXi ISO I downloaded. I selected the local hard disk as the location to install and let it complete. Once installed, I used the ESXi configuration to set the password, IP, gateway, and hostname for each server. I then installed the Virtual Infrastructure client on a windows server and used that to manage each of the R805 servers individually.

For shared storage I used a PowerVault MD3000i iSCSI storage array. I enabled the iSCSI software initiator on each server and discovered the the MD3000i. On server A, I created a VMFS partition and created a new VM called VMTest1. I installed Windows Server 2008 64-bit Enterprise Edition. After install completed I shutdown the VM on server A. I then went to server B and rescanned the storage adapters. It found the new VMFS partition on the shared iSCSI LUN. I created a VM using the same settings as I had on Server A including the same virutal hard disk file. I then boot the VM successfully.

The most interesting part of this test was to verify that the cluster file system of VMFS was still working without Virtual Center in the picture. So with the VM still running on Server B, I tried to start it on Server A - and I got an error message that the file was in use by another server. This was great because it showed that it would not be possible to run the VM at the same time on both servers.

We will chat more about this on today's chat - ESX and ESXi.

Todd

We are having a chat this afternoon, at 3 pm central time, that is officially titled ESXi Features and Licensing. I think that I should have named it The Day After with ESXi because that more clearly gets to the questions that I people are having. Many have downloaded the newly free ESXi in the last month. Installation and creating an initial VM are usually the first steps that are taken. Then you go home, sleep on it, and start to really evaluate how to use ESXi the next day.

Initially a single server running ESXi maybe all that is needed. Questions about how it works and what it can do will need to be answered. This is the Features part of the chat today. Some will arrive at the conclusion that they may need some of the features that are included with VMware's Virtual Infrastructure which includes things like live migration, high availability, load balancing, and centralized management. There are actually lots of options in terms of licensing to upgrade the free ESXi into Virtual Infrastructure. This is the Licensing part of today's chat.

Todd

Jose Maria, a Dell virtualization guru and member of delltechcenter, has his own virtualizaiton blog that he writes en espanol. Even after taking 2 years in high school and 2 more years in college I still need lots of hand waving and gesturing to get by when attempting to speak Spanish. Jose Maria is a native speaker who does very well with English as well, but has started his blog to reach the large audience of technical minded Spanish speakers. He has many excellent posts and updates it several times a week with entries that disucss the latest developments in the virtualization industry. This is a great opportunity for those of you that would prefer to get some technical content en espanol and for me to brush up on my Spanish.

I've put a link to his blog right on the delltechcenter home page, giving it some international feel.

Todd

I just posted a new video to delltechcenter.com on the virtualization demos page. It is a demo of how Microsoft System Center Virtual Machine Manger 2008 (SCVMM), which is currently in beta, can be used to clone a Hyper-V virtual machine. This was fun to put together because this is the first chance I've had to work with SCVMM 2008.

There were a couple of surprising things that I discovered around the cloning demo. The first was that a VM must be prepared in advance by running sysprep (at least for windows VMs) before doing the clone. With VMware Virtual Center it is possible to have the sysprep and other customization done as part of the clone. The second surprise to me was that Windows Server 2008 includes sysprep. In the Windows\system32\sysprep directory it is there waiting to be run. And when it is run there are only a couple of very simple options to pick from. I think that it is a much easier tool to use this way. With previous versions of Windows Server the sysprep tools were on the installation cd in a deploy.cab file, which meant you had to track down the cd. I like it better just included with the OS install - although I wonder if this is a potential problem.

In addition to the Hyper-V clone demo, Scott also put up a cool demo of SCVMM managing some VMware ESX Servers and VMs. So you might want to check them both out while you are there. We're going to talk about these demos in today's chat on Hyper-V Management.

Todd

Virtual Geek has a some great performance and power consumption numbers up on his blog for virtualizing Microsoft Sharepoint. I've gotten lots of questions over the past year or so about how SharePoint would perform in a VM and this is the best data that I have seen to answer those questions. He has info about how you can download the full report from EMC's Powerlink site, but the charts and info in his blog post seem to cover the main points. The amount of savings in terms of power are so big as to be unbelievable, but after looking at the details behind them it all seems to add up. An excellent read - I recommend that you check it out.

Todd

During the weekly DellTechCenter Tech Tuesday chat a couple of weeks ago a question came up about performance monitoring with the PowerVault MD3000i iSCSI array. The initial question was how to do performance monitoring from Linux and we addressed it in a followup discussion thread. The answer at this point was to use iostat for Linux or perfmon in windows to monitor performance on each host that was attached to the MD3000i array. This lead to a follow on question which was how to monitor the performance for the entire array.

The management tool for the MD3000i, PowerVault Modular Disk Storage Manager, does not include any performance stats beyond basic iSCSI port level stats. Some investigation into the command line interface for the MD3000i revealed that there is a command to capture the performance stats on the array. Using the smcli, which is installed as part of the MDSM, it is possible to collect the performance stats for the array, controllers, and virtual disks to a csv file.

This turns out to be pretty cool and not too hard to do. Just a simple command and a little bit of spreadsheet magic and you too can produce cool performance graphs of your MD3000i. The details are posted here on delltechcenter.com.

Todd

In the past I was somewhat biased towards 2-socket servers for virtualization - which was due to the results of some testing we did. I had of course realized over the past year or so that things had changed and the 4-socket servers were now more competitive. So when I ran the series of chats on Selecting a Server for Virtualization, I decided that it was a good opportunity to re-run some of those exact same tests with the brand new R900 that had arrived in the lab. I posted the results on a TechCenter page so you can see the full results, but the short summary is that the R900 stacks up as more efficient than the 2950 we tested in the previous paper. The R900 showed 10 to 23 percent better performance per watt than the older 2950. The question that remains is how would the R900 compare to a current generation 2950?

Todd