Year: 2006

Sometimes you could need to setup particular network scenarios in virtual machines, where routers or firewalls need 4 network interface cards.
While the new VMware Server 1.0 supports 4 of them, VMware Workstation 5.5.1 (and older) does not.

The following how-to details how to add the 4th NIC to new Workstation virtual machines or existing ones.

1. Create a new virtual machine with 3 NICs as usual and power it on (even if it has no operating system inside). After few seconds you can power it off and shut down Workstation.
If you want to add 4th NIC to an existing virtual machine just jump to step 2.

2. Go to the directory where the virtual machine is stored, open the .vmx file with Notepad or any other editor tool and add following 6 lines on bottom:

ethernet3.present = “TRUE”
ethernet3.addressType = “generated”
ethernet3.generatedAddress = “00:0c:29:08:35:c3” (this will change at next boot)
ethernet3.generatedAddressOffset = “30”
ethernet3.connectionType = “bridged” (replace with anything you like)
ethernet3.vnet = “VMnet4” (replace with anything you like)

3. Save the file and power on your virtual machine.

Note: This trick will not work to add 5th network card on Workstation 5.5.1 or Server 1.0, so don’t waste time trying.
This depends on amount of PCI slot VMware defined for its virtual computer (here more details).

Remember that 4th NIC in VMware Workstation virtual machines is unsupported and effects on virtual network performances and reliability are unpredictable.
Since VMware Server already supports it we can assume the formal limit will be extended in upcoming Workstation 6.0.

The physical to virtual (P2V) migration is one of the most attracting aspect of virtualization for community, which is working with great efforts to produce new solutions.

After EZP2V, a new tool surfaces: dd2vmdk.

As the name suggests it only works with VMware infrastructures, converting a raw disk image acquired with dd unix tool in a .vmdk virtual disk.
Since dd is included in quite every Linux distribution around, you can use any liveCD to perform image acquisition (but this means it’s not able to do live migrations).

dd2vmdk approach is atypical since the tool just details how to change the disk geometry of grabbed image, instead of actual performing the operation.
The whole thing is done online, within a web application which requires an initial input to calculate resulting .vmdk disk structure. A
After recalculation dd2vmdk list commands to be launched on a Linux platform to modify disk geometry.

The tool is still in beta and definitively not easy to handle, but it’s worth to check.

Quoting from the Transitive official announcement:

Transitive™ Corporation, the leading provider of software that enables transportability of applications across multiple processor and operating system pairs, today announced that eminent computer scientist Dr. Mendel Rosenblum has joined the company’s Technical Advisory Board (TAB).
Transitive’s TAB consists of senior industry leaders and external technical experts who collaborate with and advise Transitive’s management team on strategic technology development.
…
Dr. Rosenblum is an associate professor of Computer Science and, by courtesy, Electrical Engineering at Stanford University and a co-founder and the Chief Scientist of VMware, the global leader in virtual infrastructure software for server and desktop virtualization.

Quoting from the Qlusters official announcement:

Qlusters, Inc., the leading provider of open source data center management and automation software for physical and virtualized environments, today announced openQRM plug-in support for the Xen™, VMWare™, QEMU and Linux VServer projects. Each plug-in takes advantage of openQRM’s advanced virtualization layer and provides a proven method for IT administrators to start applications either on a physical server or virtual environment without having to reconfigure file system images.
…
To support each virtualization technology, the openQRM platform comes pre-configured with a generic, logical layer called the “partition engine.” This partition engine provides a virtualized server resource that is used the same way as a physical system with no configuration changes required on the application’s file system-images. Plug-ins for each specific virtualization solution can then be added into the partition-engine making it easy to support the creation, management and administration of any type of server under management…

Quoting from the EqualLogic official announcement:

EqualLogic, the leading provider of enterprise-class iSCSI storage area network (SAN) solutions, today announced that it has joined the VMware Technology Alliances Program as part of an ongoing effort to fully integrate EqualLogic PS Series iSCSI SAN solutions with VMware’s suite of virtual infrastructure software-including the new iSCSI-enabled VMware ESX Server, which is part of VMware Infrastructure 3.
…
In addition to participation in the VMware Technology Alliances Program, EqualLogic and VMware have signed a cooperative support agreement to ensure that customers experience seamless support for their deployment of the two companies’ combined product sets. The agreement provides VMware and EqualLogic direct access to each other’s support engineering resources to jointly support customers who have deployed VMware virtual environments with a PS Series SAN.

Despite company efforts in reaching a standardization in hypervisors and its presence at the conference, VMware has not even been mentioned at Ottawa Linux Symposium keynote, presented by Jonathan Corbet, co-founder of Linux Weekly News.

Colbert talked touched several topics including virtualization, referring to Xen and User Mode Linux (UML) as the big players in his slides, and mentioning Linux-VServer, OpenVZ and BSD Jail as other efforts.
VMware fallen in the various proprietary offerings entry.

It’s a bad result considering last week VMware attack against controverse Microsoft-XenSource agreement.

IBM released a very interesting whitepaper about VMware technologies:

This paper discusses a performance optimization in the VMware ESX Server 2.5.3 and ESX Server 3.0 scheduler that can benefit Intel® processor-based IBM® NUMA-class servers with hyperthreading enabled. This paper describes the test environment and presents the results obtained with and without the performance optimization.

The information presented is based on our experience with setting up and running VMware ESX Server configurations in the IBM System x™ Performance Laboratory.

Read the it at source.

The new vizioncore product called esxMigrator, discovered two weeks ago by virtualization.info, has now reached beta 2 and is expected to be available within 2nd August.

The product promise to simplify and speed up upgrading task when moving from a VMware ESX Server 2.x to the new Infrastructure 3.

The virtualization.info Virtualization Industry Roadmap has been updated accordingly.

Lastest version of acclaimed PlateSpin phisical to virtual (P2V) migration tool, PowerConvert, is expected to be released in the second week of August as an official email from the company states.

The new release will permit to do live (no virtual machine shutdown) virtual to virtual (V2V) from a VMware ESX Server 2.x to the new Infrastructure 3 and will support the new VMware Server 1.0

The virtualization.info Virtualization Industry Roadmap has been updated accordingly.

The Rational Guide to Scripting Microsoft Virtual Server 2005 is a great broadening of topics covered in The Rational Guide to Managing Microsoft Virtual Server 2005 from the unique perspective of scripting.

Written by the same author, its biggest win is approaching automation problems following topics arrangement of its twin book.

Chapters 1 and 2 cover basic concepts of server virtualization and product architecture itself, listing available languages for manipulating Virtual Server 2005 COM interfaces among VBScript, VB.NET or C#, VB 6.0.

Chapter 3 and 4 approaches scripting rules respectively with VBScript and .NET (with examples both in VB.NET and C#), detailing COM objects model and suggesting configuration of Microsoft Visual Studio 2005 as development environment.

Chapter 5 teaches how to obtain most informations from virtual machines accessing Windows Management Instrumentation (WMI) interface, exploring VirtualMachines and VirtualNetwork classes, and how to monitor performances accessing Windows Performance Monitor counters.

Chapters from 6 to 8 are the heart of the book, detailing automated management of virtual machines (Chapter 6), virtual disks (Chapter 7) and virtual networks (Chapter 8).
Frequent recaps to Virtual Server architecture are disseminated in these chapters, to refresh product knowledge.

Chapter 9 and 10 are dedicated to advanced concept like triggering scripts by specific virtual infrastructure events and automating Virtual Management Remote Console (VMRC) server and client operations.

Source code of book scripts is available only online for registered readers and it’s really worth to download to avoid rewriting all the code and speedup learning.

Conclusion
In a young virtualization world where out-of-the-box datacenter automation is still a big missing but a most wanted capability this book helps you approaching Virtual Server 2005 scripting fundamentals to fill the gap.

While inappropriate for programming novices, this book, unique in its kind, is highly suggested as starting point for automating complex tasks and streamlining virtual infrastructure management. And considering both Virtual Server 2005 and Visual Studio 2005 are free, it’s a cheap one to become more productive in small time.