Distributed Resource Scheduler or DRS is an add-on feature of VI 3 infrastructure that is managed by Virtual Center. DRS allow balancing the CPU and memory resources of the virtual machines or VMs and the other ESX servers in the cluster.

DRS helps to balance the CPU and memory of its cluster members based on the configured resource pool policies such as shares, reservations and limits. The hosts and VMs are continuously monitored by the virtual center. Based on the configuration, if there is any imbalance of resources, the VMs are moved across the hosts in the DRS cluster.

The placement of VMs across the cluster can be configured based on:

1. Affinity and anti-affinity rules – Rules that define which VMs can run together (affinity) and cannot run together(anti-affinity) in any given host. A perfect example for anti-affinity would be placement of a SQL server and Exchange server. At any point of time, you don’t want to place both the servers in the same host.
2. VMotion compatibility – VMotion has it’s own set of requirements to move the VMs across the hosts. For example, if a VM that has a local network (not connected to any physical adapter) cannot be moved using VMotion.

Based on the environment and needs DRS automation can be set to the following levels:

1. Manual – DRS only provides recommendation on placing the VMs. Manual action is required to place them on recommended hosts
2. Partially automated – During VM power-on, they will be placed on the DRS recommended hosts. VM migrations caused by resource imbalance will be recommended by DRS but won’t be moved automatically
3. Fully automated – DRS automatically places the VM during power-on also during resource imbalance on the DRS recommended hosts. The migration threshold level can also be set with this level between conservative and aggressive using a slide bar.

Few factors to consider about DRS:

• There can be up to 32 hosts per DRS cluster.
• It’s recommended to use combination of DRS automation levels based on the critical nature of VMs. To accomplish this, the cluster level DRS automation can be overridden by the VM level automation setting.
• In the manual and partial automation level, it is important to pay attention to the number of stars on the recommendation. A 5-star recommendation should always be considered and applied.
• Swap file location for the VMs is configurable in the DRS cluster and it is recommended to keep the swap file in the same directory in the VMFS datastore for performance reasons. Choosing to keep the swap file of the VM in the datastore based on the host setting will result in a poor VMotion performance during a resource imbalance.

Resources:

More details about DRS cluster can be found here:

http://www.vmware.com/pdf/vmware_drs_wp.pdf

http://pubs.vmware.com/vi301/resmgmt/wwhelp/wwhimpl/common/html/wwhelp.htm?context=resmgmt&file=vc_cluster_concepts.6.6.html

Recently, a customer had asked whether we can assist them in moving to a Hyper-V solution from Virtual Server.  There is a VMC to Hyper-V Import tool that you can use in order to achieve this.  You can read more on it at this technet post:

http://technet.microsoft.com/en-us/magazine/2009.03.utilityspotlight.aspx

Cloud Computing – A buzz word that is frequently murmured everywhere by the IT folks or even by the non-IT people recently. This post describes what it really means to the beginners those who are interested or looking into cloud computing.

Cloud computing is a computing model where the infrastructure and the application (even the platform) is offered as a service over the Internet. The infrastructure cloud could include servers and storage and the application cloud includes various applications.  This is a major shift in the industry in the way that software and services are going to be delivered in the future.  At the center of cloud computing is the virtualization technology which we have talked about a lot in the past.  Virtualization technology through resource pooling provides the engine that drives much of the cloud.  Companies such as Google, Amazon have already started using their server farms and offering services and Microsoft is not far behind with their Azure service.

Even though cloud computing can be classified into many different types, the major ones are Public clouds, Private clouds, Hybrid clouds and Community Clouds.

Public clouds – As the name suggests, it is usually offered by a company who has invested a lot building their datacenter and offering a part of its infrastructure and platform for a monthly fee. Amazon, Terremark, RackSpace and Google are great examples of public clouds. And Microsoft is joining the game as well with their Azure services platform.

Private clouds – This is something that enterprises build by themselves to be utilized across their organization. This allows them to consolidate their servers (and storage) as a single entity that can be offered to their different business units as needed. There is an interesting article from Network World can be found here.

Hybrid clouds – This is an emerging area of cloud computing where the private and public clouds can be integrated. There are many factors such as security and application compatibility needs to be considered in this model.

Community clouds – These are clouds that are shared by organizations having common interests.  This is similar to the IT-SSO post that we had done a couple of days ago.

Driving factors for moving towards cloud omputing:

The recent developments in the virtualization technology gave a big boost to cloud computing. There are many reasons that drive the cloud computing. Some of them are:

•    Rapid deployment of servers and applications
•    Easier scalability
•    Allowing IT to run as a cost center by running multiple datacenters as single entity which can be shared and charged back based on usage
•    Cost efficient “pay as you go/use” pricing model
•    Greener initiative due to less power consumption since the shared capacity of a virtualized cloud data center reduces the power consumption for everyone.

Apart from its benefits, there are still few concerns about the security, compliance, performance and the application compatibility with cloud computing. However, they are being addressed by the cloud vendors.

We will look into some of the cloud services in-depth in Part II

In one of our previous post, “VMWare ESX Server – Virtual machine templates“,  we had discussed the different options on creating templates.  In this post, we will cover cloning.

Cloning is an alternate method of easily deploying a virtual machine. Cloning creates an exact copy of the virtual machine. It is also possible to clone a template. There are different types of cloning options available with VMWare Converter Enterprise or VCE. VMWare Converter Enterprise is a windows application and comes with Virtual Center Server.  Here are the different options:

Cold cloning (or) offline cloning – This is the simplest form of cloning. As the name implies, the source virtual machine need to be powered off before cloned. The system to be cloned is booted with the standalone version of VCE CD. Once the user selects the destination and starts the migration, VCE creates destination virtual machine by copying the volumes from source. It also allows customizing destination OS parameter such as IP address.

Hot cloning – This cloning can be done while the source virtual machine is powered on. Using the converter plug-in VI client, the whole process can be automated without user intervention.

Cloning can be further classified into ‘Remote cloning’ and ‘Local cloning’ based on how the cloning is initiated. If the cloning is done using the converter plug-in on VI client by installing remote agent on the source virtual machine, it is called ‘Remote cloning’. If the cloning is done using the stand alone version of VCE CD, it is called ‘Local cloning’.

More information about VMWare cloning can be found at the following VMWare site:

http://www.vmware.com/support/ws55/doc/ws_clone_overview.html

There are some good resource management features available with VMWare that allows to effectively utilizing the CPU and memory of ESX physical host. We will discuss about them in the next post in the VMWare category.

Among the many best features of ESX server, virtual machine templates is one of the time saving features that system administrators can rely on to deploy servers at ease.

A template is a ‘pristine’ image of a server that can be used to create and provision new virtual servers on the fly in the ESX server environment. The template includes an operating system, its configuration and other set of applications installed on a typical server.  There are four options related to templates in Virtual Center when you right click on a Virtual Machine or template that a system administrator needs to be aware of. Each option performs a different task and it is generally confusing to the new ESX server administrators. This post explains those options.

The options are as follows:

1. Migrate – Selecting this option moves virtual machine from on ESX server host to another host. Without VMotion installed, this is a cold migration process that requires the virtual machine to be powered off before being migrated.
2. Clone – It creates an exact copy of a virtual machine. Guest OS customization is required before powering on the cloned virtual machine in order to avoid SID and IP address conflicts in Windows operating system.
3. Clone to Template – This option keeps the original virtual machine and creates a template.
4. Convert to Template – This option converts the selected virtual machine to template and the original virtual machine disappears

Once a template is created, using Virtual Center it can be used in the following ways:

1. Cloned to create a new template
2. Converted to a virtual machine (The template disappears once the virtual machine is generated)
3. Deploy virtual machine – This wizard driven option allows to keep the original template and offers guest OS customizations (using Sysprep) to create a new virtual machine

Few facts about templates:

•    VMWare Virtual Center is needed to create a template. It cannot be created with a stand alone ESX server
•    Never power-on a template
•    Templates can be used to deploy virtual machines across different ESX servers hosts managed by the Virtual Center

There are a few interesting cloning options in VMWare provided by VMWare converter.   We will discuss those in the next post.

References:

• More information about VMWare templates and the best practices can be found in the following VMWare document:  http://www.vmware.com/pdf/vc_2_templates_usage_best_practices_wp.pdf

SQL Magazine had an excellent article by Jeff James on virtualization in the SQL Server world…also read the Myths and mis-conceptions article from the same author. Here are the links:

Virtualizing SQL Server – here.

Virtualization Myths and Mis-conceptions – here.

Microsoft seems to be teaming up with Red Hat as well as Citrix.  And VMWare seems to be teaming up with Intel.

A lot has been said and written about virtualization and a lot of companies are finding benefits adopting it in their IT infrastructure or the data centers that they might be hosting their applications with.  At Decipher, we have been supporting our clients with their virtualization needs specifically with the VMWare and Hyper-V solutions and have also blogged about it before.  Recently, while making a presentation to a potential client, one of the questions that was asked by the Operations Manager was to give a brief overview of virtualization and the different types of virtualization.  We are sharing that with our readers as well here.

Basically, virtualization is dis-associating the tight bond between software and hardware.  In a traditional world, one would use one server for their Oracle instance, one for a Microsoft instance, yet another one to host their Exchange Server and so on for the other applications that they have to support.  At times, you would be able to consolidate them assuming that the OS, providers and the different vendor requirements are common.  What this ends up doing is give you a tons of servers to manage, which consume more power, waste the resources on these boxes because some of these servers would be under-utilized (CPU, Memory wise).  Not only that, if you need to add more capacity you would either try to scale up (more memory or additional processors if possible) or scale out (more servers).  By being able to dis-associate the bond between the software and the hardware one can then use the same hardware to serve up the needs of the different software servers.  So, you could be running Oracle, MS SQL Server, Exchange, Great Plans, Dynamics CRM, etc. all on the same hardware.  By doing this, it is also possible to run different operating systems so I could run MS SQL Server 2008 on Windows 2008 Server and at the same time run, Oracle on Linux all running on the same hardware.  By doing this, the resources will be better utilized and also allow us to easy add another VM on the same hardware assuming that we have the capacity.  There are additional benefits as well like making use of the vmotion capabilities of VMWare etc..

So, that was a 50,000 ft. overview of virtualization.  So, how does it really work?  What breaks the bond between the software and the hardware?  We are going to talk only about VMWare’s ESX Server 3.5 and MSFT’s Hyper-V since the older offerings from both these vendors had a different architecture.  Both of these are based off hypervisor based architecture.   A hypervisor which is also sometimes called a VMM (Virtual Machine Monitor) is essentially a hardware platform virtualization software using which one can run different OS on the same hardware at the same time.  Hypervisor has access to the physical host hardware.  Even though both Microsoft’s solution as well as the VMWare solution are based on the same hypervisor based technology, there are differences in how the implementation is done.  We won’t cover all those differences here but will provide some links

Now, what are the different types of virtualization available and what are the dominant players in the market that provide a solution for it.  There are essentially three different types:

1) Hardware Virtualization:

This is the most common and is used in IT departments in a company as well as in the data centers.  The server’s hardware is virtualized thus allowing us to be able to run different OS and different applications simultaneously on the same hardware.  This allows us to do server consolidation.  And the benefits are obvious (only listing the critical ones here…and less cost is a major advantage across all of these):

a) Less number of servers required for the same number of applications.
b) Less power consumption.
c) Less maintenance overhead for the IT staff.
d) More resource utilization.
e) Easier (and faster) to add more capacity.
f)  Patch management and upgrades become easier.
g) DRP (Disaster Recovery Planning) becomes easier.  Without any interruption to the service, one can backup and even migrate entire virtual environments.

2) Desktop Virtualization:

We have gotten into the habit of calling this VDI since that is the term that VMWare uses for desktop virtualization.  VDI stands for virtual desktop infrastructure.  But this is not limited to just VMWare.  Citrix systems has a similar offering called the XenDesktop.  What this means is that your end user’s computer’s data – their OS, their applications, their downloads, their preferences etc. are all stored in a VM in a hosted environment which could be hosted either by the company’s IT in-house or hosted in a data center.  Som the VMs are then managed in one single place for all the users in a department/company and the computing environment is delivered remotely to the end users.  The one reason why the adoption has been a bit slow on this front is because unlike server consolidation (hardware virtualization), desktop virtualization requires working across a lot of different organizations within the company and it impacts the end users a lot more during the stages of putting the plan in place and executing it.  Benefits are obvious:

1) Easier upgrades and patch management.
2) IT Desktop support process becomes much more easier.
3) You can easily add more users as your organization grows and provisioning of new applications and VMs takes minutes and not days/weeks.
4) Better resource utilization and less power consumption.
5) Easier recovery management.

3) Storage Virtualization:

So, consolidating servers as well as the desktops is all great but what happens to the storage requirements then?  Won’t the storage requirements also grow by leaps and bounds?  This is the next question that you are going to get from your clients – internal or external.   This also means that since everything is in one place, one also needs to have a proper plan for disaster recovery and business continuity.   So what does storage virtualization mean then?  It means we would then need to make multiple storage devices appear as a common shared media.  A proper back-up and restore strategy needs to be formed as well then and a proper DRP needs to be done – both local and site failures need to be accounted for.  We will present a more detailed DRP analysis in one of our whitepapers that we are working on.

Major Players:

So, who are the major players in this market?  If you have been following the virtualization world, you already know that VMWare, Citrix and Microsoft are the three leading players in this market right now.  There are other players as well who have entered the market – Red Hat, HP, IBM, Oracle and Virtual Iron Software are some of the other names.  For end consumers like us, the more the competition in the virtualization industry, the more affordable it is going to be for us.  Microsoft’s solution (Hyper-V) is free but that solution is not as robust as the VMWare offering.  A lot has been already said about comparing the two.  Here are some resources to help you with that comparison:

a) VMWare’s Presentation: Here.
b) WindowsIT Pro comparison (this requires a subscription to read the article): Here.
c) Good Discussion @ Virtualization.info: Here.
d) ITComparison.com’s Analysis: Here.
e) Michael Otey’s article: Here.

Do note that we do not recommend using virtualization for your database servers IF you have a highly transactional system or a datawarehousing/data mining system that requires heavy IO activity.  Read more over here.  If you have applications which do not tax the database that much and the IO is relatively less, virtualization works perfectly fine.  This is based on our research as well as limited tests that we have done in our lab & client sites.

In an attempt to counter the open source virtualization offerings, VMWare has released their open source virtualization client.  You can read more on this here.  The VMWare view open client allows folks in a company to access their desktops from any device from anywhere since the hosting of the virtualized desktops is done either in-house or at the data centers.  This does pose a big threat to MSFT’s Hyper-V solution since this gives an also gives the businesses an opportunity to slow down PC purchases regardless of their increasing head count.

We have covered virtualization quite a bit in our blog posts before. Saw this blog post at MSDN regarding the support for SQL Server in a virtual environment which is something that you should be familiar with if you are (or are planning to) running your SQL Server environments in a virtual environment. In addition, as we are playing with Hyper-V nowadays and were researching our choices for SQL Server 2008, here are two sites that you might want to bookmark for your reading purposes:

1) The SQL Server 2008 virtualization section – this has a lot of very good links and resources.

2) Whitepaper by the SQLCAT team – running SQL Server 2008 in a Hyper-V environment.