VMGuru.nl – I choose (a virtual) life!

At a client site we came upon a problem with Windows 2003 VM’s. They would get low network performance when we moved them to a newly formed ESX cluster consisting of HP 460c G6 blades. In some cases logging on to the server with a remote session took about 20 minutes.

As I mentioned this only occurred when we moved a VM to the new cluster, but also VM’s that where newly installed would get the same problem when running on the new cluster. As we are using Altiris to install and configure new VM’s a colleague decided to install a new VM by going through the steps manually which normally would be done by Altiris and found out that after the activation of a security template the performance dropped significantly.

(more…)

We recently received an update about the new stuff that is coming from Dell EqualLogic. Because a lot the information was under NDA so I can’t go into details or give you release dates.

What’s here already?

Last year Dell released the 10GB connections for the PS series. This 10GB connection speeds are mostly done because the marketing asked for it, because the customers asked for it. There are rarely used cases where the 10GB connections make sense but in 95% of all cases it doesn’t give you extra performance.

The bottleneck is often not the connection speed but the spindle speeds limits of the disk themselves. Because of the adaptive load balancing of the PS series the 1GB connections are used very efficiently. When used with the MPIO drivers from Dell EqualLogic you can gain a lot extra performance from the network connections.

So what is coming this year:

(more…)

Well, it’s done, as far as we’re concerned. The migration to new hardware is a fact. You might see a significant speedup in the site (we do). We’re very happy with the results, allthough we had some issues, but then again, which migration hasn’t?

We made some pics during the swap. The quality is pretty poor but you can’t expect much from a mobile phone. Here they come! (more…)

Good news! Our brand new server is finally ready for production! Next Friday we will replace the old workhorse with a brand new one, with many thanks to our friends @ Dell (a big hurray and thank you to you all over there!) Our new infra will, of course, be fully virtual. Since we have more processing power AND more RAM, our site should benefit a lot in terms of speed and stability.

We hope that service will hardly be interrupted but as with all migrations, you never really know until it’s over

Work starts Friday at 10:00AM CET and ends when it’s ready

Update 15:41 CET+1 (AJ): Basic functionality of the site is up again. We’re still doing some tweaking during the next couple of days. If you get an 404 error, just try again.

Last week I have been struggling with  the installation of a vSphere 4 infrastructure on Dell hardware at a Belgium client site.

I have done many many many VMware installations and encountered my fair share of issues but apart from the HP USB sticks the hardware never gave me this much trouble.

It all started with a very difficult BIOS/firmware upgrade which, after various downloads and trials, ended with an old-school DOS boot USB and a DOS based BIOS update. Real 1980’s stuff.

With this fixed I installed all ESX hosts and left for the hotel, ready to start the configuration the next day. However, when I started with the first ESX host and wanted to configure the network, I noticed that I only had eight NICs when I should have had twelve. We use Dell PowerEdge R805 servers with two Intel quad port 82576 Gigabit Ethernet Adapters, the first card was already in the server, the second card we added just before the installation.

(more…)

When designing a virtual infrastructure an important bit in the design is the storage infrastructure also called the Storage Area Network (SAN). In a SAN based on iSCSI we often use Cisco 3750 switches, but when you are going to select the right Cisco 3750 for the job the fun starts. You will be dazzled by the amount of different product numbers and will be busy deciphering the product code.

The product code for a Cisco 3750 switch is build up like this:

WS-C3750a-xxbc-dee

WS stands for Switch
C stands for Catalyst series
3750 stands for the 3750 product line

a >> blank, G, E
blank = classic 3750 switch, 6.5 or 13.1 mpps forwarding rate
G = all ports are gigabit, 35 or 38 mpps forwarding rate
E = enterprise line, 65.5 or 101.2 mpps forwarding rate

xx >> 12, 16, 24, 48
12 = 12 Ethernet ports
16 = 16 Ethernet ports
24 = 24 Ethernet ports
48 = 48 Ethernet ports

b >> T, P, F, D, W
T = Ethernet ports
P = Power over Ethernet

(more…)

Are you having trouble consolidating servers?

Having troubles achieving the consolidation ratios you wanted?

Maybe Intel has got the solution! Yesterday, Intel has presented a 48 core CPU!!

Let’s do the math, 4 virtual machines per core, 48 cores that’s 192 virtual machines per CPU. Take a dual socket server, add two of these CPUs and you have 384 virtual machines per host. That was the good news, the bad news is that this 48 core CPU is only available to research institutes.

The CPU which is developed and produced in Europe has 48 cores which are grouped in pairs of 24 all with their own L2 cache. The 45nm CPU contains 1.3 billion transistors on area of 567 square millimeter. Communications between the core pairs is handled by a message buffer and router and a 256GB/s mesh network. Six pairs share one memory controller and can maximally address 32GB of memory. Energy consumption is between 25 and 125W.

As mentioned before this 48 core CPU is only available to research institutes but some of the techniques used will eventually be used in CPU we use for our virtualization magic. Let’s wait and see.

On a regular basis we have info sessions with our most important vendors.  Last week we had a session with HP to tell us more about virtualization in their hardware products. The session was especially targeted at Flex-10. Flex-10 is the way how HP breaks s a 2 x 10Gb Ethernet pipe into a flexible, easy to change, smaller Ethernet ports.

Why is this so important for us virtual friends? Of course it is a huge cost saver not only in hardware but also in management of the environment but the most important thing is that it opens up a lot of new virtual design opportunities.

.

One of the coolest things is that we now can make a design for up to 4 blade chassis with each physical 16 server blades and let’s say 320 virtual servers where all the traffic between the servers never leaves the blade chassis. It is all handled with the blade chassis. Also all of the vSphere traffic like VMotion and service console can be handled within the chassis at 10Gb speeds.

(more…)

Today Emulex announced the general availability of their new OneConnect Universal Converged Network Adapter (UCNA) Product Family, specifically the OCe10102-I 10Gb/s Ethernet iSCSI Adapter and OCe10102-N 10Gb/s Ethernet Adapter are very interesting.

The Emulex OneConnect UCNA is a single-chip, high performance 10Gb/s Ethernet adapter that provides connectivity for network and storage traffic over one multi-function server adapter. Unlike first generation CNAs that only provide FCoE convergence, the Emulex OneConnect UCNA technology provides optimized performance for TCP/IP, FCoE and iSCSI protocols.

The OneConnect UCNA comes in three different versions (N, I and F):

• OCe10102-N: 10Gb Ethernet Adapter;
• OCe10102-I: 10Gb iSCSI Adapter;
• OCe10102-F: 10Gb FCoE  Adapter.

(more…)

Now that the new version of VMware View is coming shortly I was investigating the new possibilities with our thin client providers. Wyse Technology which has a nice solution already for the View 3.1 platform now has a new Zero Client which is specially made for View 4.0.

What makes this Zero Client so special? This model support PCoIP. PCoIP is the answer from VMware to make a great user experience possible with VMware View. VMware partnered with Teradici to introduce the PCoIP technology into View.  The P20 not only uses the software implementation but the P20 also contains the hardware chip that Teradici uses in the PCI Card solution. Tested by Brain Madden a couple of month ago.  Also see this article at BrianMadden . Imagine the possibilities this solution can offer in your VDI environment.

When the Wyse P20 comes to EMEA is not yet clear, but I hope to find out more next month. We at VMGuru.nl will also try to get our hands on a P20 model so we can blog our experiences.

More info on:  http://www.wyse.com/products/hardware/zeroclients/P20/index.asp

It is becoming a sequel, the datacenter VMware has build for this weeks VMworld 2009 at the Moscone Center in San Francisco.

In addition to our two previous articles (art1, art2), today we found two very nice videos loaded with tons of techno porno!

The first video shows the VMware team building the complete datacenter on-site at the Moscone Center. During the video footage the awesome numbers representing this huge infrastructure run by.

In short? 28 racks containing 776 ESX servers which provide the infrastructure with 37TB of memory, 6.208 CPU cores and 348TB storage which uses 528KW electricity and is servicing 37.248 virtual machines. You will probably never find such an infrastructure anywhere in the world, at least I know I won’t.

(more…)

Last week Iomega  announced a new NAS appliance called the Iomega StorCenter ix4-200d. It is based on  EMC storage technologies and contains four SATA-300 drivebays. With a storage capacity of 2, 4 and 8 TB it is the ideal NAS appliance for small businesses, distributed offices and home office networks that require advanced data storage and protection features without the need for a dedicated IT staff.

By default the Iomega StorCenter ix4-200d is configured as a RAID 5 array but RAID 10 or JBOD is also possible. Free disk space can be monitored on the front LCD display.

The Iomega StorCenter ix4-200d is equipped with two gigabit ethernet ports which support link aggregation, network failover and subnetting. It also supports various protocols like iSCSI, NFS, DLNA, uPnP, Bonjour, SNMP, FTP, HTTP and Samba so it can be used with Mac-, Windows- or Linux hosts.

(more…)

Scouting the Internet on information about the new G6 evolution of HP servers I ran into a valuable article from Martijn Lohmeijer’s blog.

In a lessons learned article he warns about the new memory management in the new HP ProLiant DL360/380 G6.

We all know what happens when new hardware is delivered (at least I do). Techies as we are, we start unpacking the servers, flip the cover and inspect the new Intel Nehalem CPUs, fiber channel HBAs and SSDs murmuring about GHzs, GB, IOPS and FSBs with some additional ‘oh’, ‘ah’ and ‘cools’. Then put in the memory, disks and other toys we ordered and rush it to the server room or test area, plug it in, install an operating system or hypervisor. We then find out that the server is not performing as expected and start screwing up BIOS settings and pulling out components.

All you should have done after unpacking your new present is reading the manual and the quick specs of the server (which we (as techies (and men) never do).

The following is a enumeration of the most important items from the quick specs of the HP ProLiant DL360 G6.

DDR3 memory population guidelines

Some DIMM installation guidelines are summarized below:

• DIMM slots that are white should be populated first
• Do not mix Unbuffered memory (UDIMMs) with Registered memory (RDIMMs)
• If only one processor is installed in a 2CPU system, only half of the DIMM slots are available
• Do not install DIMMs if the corresponding processor is not installed
• To maximize performance, balance the total memory capacity between all installed processors
• You can only install two quad-rank DIMMs per channel
• You can only install two UDIMMs per channel; if available, the third slot in the channel must remain empty

A full list can be found in Martijn’s article or on the HP website.

After some hard pushing and nudging with Oracle sales the last couple of months I almost became a Oracle licensing guru. Not what I had in mind and was aiming for to be honest. While completing some business cases about virtualization for several customers, Oracle products became a hot issue again.

The Oracle Soap

I advised the customer to be careful about mentioning that they were striving for virtualizing the Oracle servers on VMware. The Oracle account manager could smell blood and would jump on the band wagon to let them pay for their attempt to make the infrastructure flexible and ready and supportive for a fast changing business. So the first thing what happened when the Oracle account manager heard the word virtualization was that he mentioned to the customer that it would cost at least 200K Oracle licensing costs even without knowing what the customer was pursuing.

(more…)

I have had several customers asking me for advice on what to do with new ESX hosts who should be joined into the VMware cluster, but after adding to the cluster problems with vMotion arose. It just didn’t work anymore because of some minor CPU differences.

One of the customers had bought the exact same HP DL380 model with same product number and revision with the same type of Intel CPUs inside. But unfortunately the stepping on the old CPUs is 6 and on the new ones its 10. The HP machines contain Xeon CPUs from Intel type 5450 3 Ghz stepping 6 (existing) and stepping 10 (new).

I asked them the following questions so I could give them advice on what to do next:

Question: How many new ESX hosts are you adding to the cluster?
Impact: If you are just adding 1 or 2 ESX hosts as extra capacity it is good to look into reforming the VMware cluster to an EVC cluster., because the more ESX hosts (up to 32) in a DRS cluster the better DRS can do its job. If you have more than 2 hosts to add to the cluster it can be a solution to build a dedicated cluster with the new ESX hosts.

Question: Do you need the new features added to the CPUs or do you just need more power in the VMware cluster?
Impact: If you look at the latest range of CPUs it can make a total difference of up to 25% speed, because of new added features.

Question: What are the plans for the future with the clusters and do you suspect significant growth?
Impact: If you suspect significant growth it can be useful to build up a new ESX cluster with new functionality but always weigh carefully the pros and the cons.

Question: Are the used servers and CPU’s capable of switching on the VT or AMD-V option and can the XD or NX bit be enabled in the BIOS? (Intel markets the feature as the XD bit, for eXecute Disable. AMD uses the name Enhanced Virus Protection.)
Impact: If the machine and the CPUs are capable, you can start using an VMware EVC cluster.

After answering the above questions I recommended some clients to build up an EVC cluster in vCenter Server, most answers I get after suggesting such a move are:

“uhhh build an EVC cluster, what’s that?”

(more…)

A few weeks ago I attended the ‘What’s new in VMware vSphere 4‘ course given by Eric Sloof. Eric pointed us to a very nice and cheap shared storage solution from Iomega which he also uses in his home lab, the Iomega Storcenter ix-series.

I have been scouting the local hardware vendors for the Iomega ix2 but I could decide to go for the cheaper 1TB or the more expensive 2TB version. Today I was browsing through the VMware HCL when I remember the dilemma and was wondering if VMware even knew/supported the little Iomega storage box. And what do you know ………… the Iomega ix-series is on the VMware HCL!

The Iomega StorCenter ix2 and the ix4-100 are supported for all versions of VMware ESX 3.5 and the larger rack mountable ix4-200r is even supported on ESX 4.0 as NAS or iSCSI.

So I will be heading for the shops to get my own home shared storage solution.

The Iomega StorCenter ix is available in the following configurations:

With the launch of VMware vSphere this week there are some new features to play with. One of the most interesting features, I think, is VMware FT (Fault Tolerance). Implementation of this features introduces some new caveats, in the case of FT you can not use all CPU’s and you can only apply FT on virtual machines with 1 vCPU. 1vCPU limits the implementations but is very easy to check in contrast to finding more compatible CPUs.

To check which CPUs are compatibel Gabrie van Zanten from Gabes Virtual World published a great article with a matrix of compatibel and incompatibel CPUs.

You can check the list and read the article here.

Update April 22th 2009: Today the rumor mentioned below was confirmed during the vSphere 4 launch.

Rumor in the VMware community has it that the next version of VMware ESX, vSphere, has very limited support for the VMware management and automation suite. vSphere will of course be available with vCenter for basic management and automation tasks but the rest of VMware’s great product suite is not supported in the initial release of vSphere.

So there won’t be support for:

• VMware vCenter Site Recovery Manager
• VMware vCenter Lab Manager
• VMware vCenter Lifecycle Manager
• VMware vCenter Stage Manager
• VMware View

In my opinion this is a mistake. I understand the pressure is on and VMware should  release a new version fast to challenge the competition. But releasing a product which does not support the majority of your former product suite is not a clever move.

From what I understand this is also a discussion within VMware so let’s hope VMware gets wiser and postpones the release or add support for the VMware Management & Automation suite soon.

Today HP announced the release of the new ProLiant G6 server line.

After visiting an NDA session with HP on VMworld Europe 2009 a lot of new G6 ProLiant servers were shown. Nice new hardware build up for even better supporting virtualization in common.

All models are made energy efficient and ready for supporting virtualization. Also other things improved like support for the Flex-10 technology, Intel Xeon 5500 CPU’s (Nehalem) which can give 25% more power for virtual workloads and more and more memory banks to fill.

(more…)

Today Dell EqualLogic introduced their new PS6000 series iSCSI SANs. The new models in this serie are the PS6000E, PS6000X, PS6000XV, PS6000S and the PS6500E.

One of the most interesting models is the PS6000S. This PS6000S is an enterprise-class iSCSI SAN with SSD technology and is offering an affordable and easy means to bring the low latency performance benefits of SSD technology. The PS6000S provides higher IOPS and as little as 1/3rd the latency of 15k rpm disk arrays, ideal for challenging applications, such as virtual desktop implementations such as VMware View with linked clone technology. The Solid State Drives (SSD) provide low-latency and high-IOPS.

(more…)

Still a lot of people hesitate to make vCenter (was VirtualCenter) virtual. They have a lot of reasons for that.

Let me sum up the most common reasons people mention when they don’t want to virtualize vCenter. After that I will (try to) explain why you should virtualize vCenter in my opinion.

(more…)

Yesterday AJ and me have been attending a seminar from Sun Microsystems about the Sunray. The Sunray is the next step in moving more processes into the backend and less hardware at the frontend.

After a little introduction of the Sunray, Sible Schöne of the Climate bureau was given the word. He told a long, but very interesting story about the climate changes on earth. In the end he basicly told us that in order to bring down power consumption the Sunray would be a great first step to do so. The Sunray in itself only uses 4W and there are options to have the Sunray draw its power from power-over-ethernet. (more…)

Today we got a server from a company the client took over. It was an old server which was used to store scans from a multfunctional printer/scanner. This server needed to be P2V-ed to be used as an archive. The first shock was that it had 2 dongels connected. Luckily these dongels were for the licensing of the multifunctional which we no longer use. Then we found out the old hardware was far from stable, a cold clone resulted in an error indicating hardware trouble. So a collegue swapped all possible hardware in an attempt to create a functioning server which would live just long enough to P2V it. (more…)

Today me and AJ went to HP to get more information on HP c-Class blades. Last week we borrowed a C3000 enclosure with a c460 and a c685 blade from HP for our Centric/VMware seminar and we were very happy with it. I have to confess that prior to these blades I had very limited experience with blades. The only blades I worked with were HPBL20p blades and it wasn’t very plesant. They needed huge amount of power and cooling and connections were very limited.

But as I said the new HP blades certainly surprised us, so we asked HP if we could come over and get information and maybe a demo of the new HP blade systems and especially how we can use them in our virtualization/VMware projects. (more…)