Home > Blogs > VMware Consulting Blog > Tag Archives: Horizon

Tag Archives: Horizon

VMware Horizon 7: Instant Clone Pools Delete VMs on Restart

by Joe Graziano

One of the observed features of Instant clone pools is that when a VM is logged off it deletes the VM so that any changes are gone. This makes troubleshooting tricky and you cannot set the do not delete at refresh like you can with Full VM and Linked Clones.

What is a View admin to do? Thanks to the folks at VMware GSS we have an option, with a caveat.

You can go into the Adam Database through ADSIEDIT and change a setting so that the VM’s do not delete, allowing you to logoff and reboot them while retaining the VM’s settings.

The caveat* is that when you change this setting you cannot edit the provisioning details or other attributes of this pool. You will have to change it back to ‘deleteonuse’ in the ADAM ADSIEDIT before you can make changes.

Continue reading

The Anatomy of an Instant Clone

By Travis Wood

If you’ve used Horizon View over the last few years, then you most likely have come across linked clones. Linked clones use a parent image, called a “replica,” that serves read requests to multiple virtual machines (VMs), and the writes in each desktop are captured on their own delta disk. Replicas can also be used to change desktop update methodologies; instead of updating every desktop, you can update the parent image and recompose the rest of the desktops.

Horizon 7 has introduced a new method of provisioning with Instant Clones. Instant Clones are similar to linked clones in that all desktops read from a replica disk and write to their own disk, but Instant Clone takes it one step further by doing the same thing with memory. Instant Clones utilize a new feature of vSphere 6 where desktop VMs are forked (that is, Instant Clones are created) off a running VM—instead of cloning a powered-off VM—which provides savings for provisioning, updates, and memory utilization.

Golden Image

With Instant Clones you start with your golden image, in a way that is similar to linked clones. The golden image is the VM you install the operating system on, then join to the domain, and install user applications on; you follow the same OS optimizations procedures you would use for Instant Clones.

When you’re done, release its IP address, shut it down, and create a snapshot. Now you are ready to create your Instant Clone desktop pool. This VM should have VM Tools installed, along with the Horizon Agent with the Instant Clone module. It is NOT possible to have the Instant Clone and Composer modules co-installed, so you will always need different snapshots if using Instant Clones and linked clones from the same golden image. Reservations can be set on the golden image and they will be copied to the Instant Clones, reducing the size of the VSwap file. It is important to note that the golden image must be on storage that’s accessible to the host you are creating your Instant Clone desktop pool on.

Template

When you create your pool, Horizon will create a template. A template is a linked clone from your golden image, created on the same datastore as the golden image. It will have the name cp-template, and will be in the folder ClonePrepInternalTemplateFolder. Template disk usage is quite small, about 60 MB. There will be an initial power-on after the template is created, but it will then shut off.

TWood_Horizon Template

Replica

Next, Horizon will create a replica, which is the same as a Linked Clone replica. It is a thin-provisioned, full clone of the template VM. This will serve as the common read disk for all of your Instant Clones, so it can be tiered onto appropriate storage through the Horizon Administrator console, the same way it is done with Linked Clones. Of course, if you are using VSAN, there is only one datastore, so tiering is done automatically. Horizon will also create a CBRC Digest file for the replica. The replica will be call cp-replica-GUID and will be in the folder ClonePrepReplicaVmFolder. The disk usage of the replica will be depend on how big your Gold Master is, but remember, it’s thin provisioned and not powered on, so you will not have VSwap functionality.

TWood_Horizon Replica

Parent

Horizon will now create the final copy of the original VM, called a parent, which will be used to fork the running VMs. The parent is created on every host in the cluster; remember, we are forking running VMs here, so every host needs to have a running VM. These will be placed on the same datastore as the desktop VMs, where there will be one per host per datastore. Because these are powered on, they have a VSwap file the size of the allocated vMEM. In addition, there will be a small delta disk to capture the writes booting the parent VM and the VMX Overhead VSwap file, but this—and the sum of the other disks—is relatively small, at about 500 MB. These will be placed in ClonePrepReplicaVmFolder.

TWood_Horizon Parent

Something you’ll notice with the parent VM is that it will use 100% of its allocated memory, causing a vCenter alarm.

TWood_vCenter Alarm

TWood_Virtual Machine Error

Instant Clones

OK! At this point, we are finally ready to fork! Horizon will create the Instant Clones based on the provisioning settings, which can be upfront or on-demand. Instant Clones will have a VSwap file equal to the size of the vMEM—minus any reservations set on the Gold Master, plus a differencing disk.

The amount of growth for the differencing disk will depend on how much is written to the local VM during the user’s session, but it is deleted on logout. When running View Planner tests, this can grow to about 500 MB, which is the same as when using View Planner for Linked Clones. The provisioning of Instant Clones will be fast! You’ll see much lower resource utilization of your vCenter Server and less IO on your disk subsystem because there is no boot storm from the VMs powering on.

TWood_vCenter Server

Conclusion

Instant Clones are a great new feature in Horizon 7 that take the concept of Linked Clones one step further. They bring the advantages of:

  • Reducing boot storms
  • Decreasing provisioning times
  • Decreasing change windows
  • Bringing savings to storage utilization

Instant Clones introduce a number of new objects: replicas, parents, and templates. It is important to understand not only how these are structured, but also their interrelationships, in order to plan your environment accordingly.


Travis is a Principal Architect in the Global Technology & Professional Services team, specializing in End User Computing.  He is also a member of the CTO Ambassadors program which connects the global field with R&D and engineering.

Managing Horizon Gold Images Across Multi-Site Deployments

By Dale Carter

One of the challenges when deploying VMware Horizon across multiple sites or data centers is how to keep your Gold/Master images in sync and how to get them from one site to another.

In this blog I will show you how you can utilize the new Content Library that is part of vSphere 6 to help manage this challenge.

There is a caveat to using the content library – it does not currently manage VM Snapshots. This blog will also show how you can work around this caveat to make the solution work for your deployments.

The following steps will show you how to create a Content Library and then use the Content Library to move your Gold/Master images between sites.

Create Your Content Library

  1. Connect to the vCenter Web Client on your home site
  2. From the home menu select Content Libraries

DCarter Gold Images 1

  1. Click Create new content library

DCarter Gold Images 2

  1. Give the library a Name, select the vCenter Server and click Next

DCarter Gold Images 3

  1. Select Local content library and check the box for Publish content library externally then click Next

DCarter Gold Images 4

  1. Select the datastore you want to save the content library in and click Next

DCarter Gold Images 5

  1. Click Finish
  2. Right-click the new Home library and click Edit Settings

DCarter Gold Images 6

  1. Click Copy Link and then OK

DCarter Gold Images 7

  1. Now connect to the web client of the remote vCenter
  2. From the home menu select Content Libraries
  3. Click Create new content library
  4. Give the library a Name, select the vCenter Server and click Next

DCarter Gold Images 8

  1. Select Subscribed content library. Then paste the link into the first library in the box and click Next

DCarter Gold Images 9

  1. Select the datastore to save the content library and click Next
  2. Click Finish

The Content Libraries are now created at each site and are ready to have content published to the library.

The next steps are to publish the Gold/Master image to the home library and then deploy that image in the remote data center.

Publishing the Gold/Master Image

The following steps will show you how to publish the Gold/Master image with the latest Snapshot to the content library.

  1. Connect to the vCenter Web Client on your home site
  2. Under VMs and Templates right-click the Gold/Master image and click Clone – Clone to Template in Library

DCarter Gold Images 10

  1. Give the new template a name, select the correct Library and click Next

DCarter Gold Images 11

The template will now be published to the Content Library and then synced to the remote library. You can speed up the sync by connecting to the remote library, clicking Actions and Synchronize Library.

DCarter Gold Images 12

Publish Template to Remote Site

The following steps will show you how to deploy the new Gold/Master image with the latest Snapshot to the remote site from the content library.

  1. Connect to the vCenter Web Client on your remote site
  2. From the home menu select Content Libraries
  3. Select the Library and click Related Objects

DCarter Gold Images 13

  1. Right-click the correct template and click New VM from This Template

DCarter Gold Images 14

  1. Confirm the name of the new VM and the location and click Next

DCarter Gold Images 15

  1. Select the correct resource and click Next

DCarter Gold Images 16

  1. Confirm and click Next
  2. Select the disk format and the datastore location and click Next

DCarter Gold Images 17

  1. Select the required Network to deploy the VM to and click Next

DCarter Gold Images 18

  1. Click Finish

The VM will now be deployed to the remote data center. However, there is one last step required before you can use Horizon to deploy new desktops – create a Snapshot for the composer to use.

  1. Right-click the newly created VM and click Snapshots – Take Snapshot

DCarter Gold Images 19

  1. Give the Snapshot a name and click OK

DCarter Gold Images 20

 

The VM is now ready to be used by Horizon to deploy desktops with the latest Gold/Master image.


Dale is a Senior Solutions Architect and member of the CTO Ambassadors. Dale focuses in the End User Compute space, where Dale has become a subject matter expert in a number of the VMware products. Dale has more than 20 years experience working in IT having started his career in Northern England before moving the Spain and finally the USA. Dale currently hold a number of certifications including VCP-DV, VCP-DT, VCAP-DTD and VCAP-DTA.

For updates you can follow Dale on twitter @vDelboy

VMware App Volumes Multi-vCenter and Multi-Site Deployments

By Dale Carter

With the release of VMware App Volumes 2.9 comes one of the most requested features so far: multi-vCenter support. With multi-vCenter support it is now possible to manage virtual desktops and AppStacks from multiple vCenter instances within the same App Volume Manager.

The following graphic shows how this works:

DCarter App Volumes

With this new feature App Volumes can now be used to support the Horizon Block and Pod architecture with just one App Volumes manager, or cluster of managers.

Now that we can support multi-vCenters, I started to wonder if this new capability could be leveraged across multiple sites to help support multiple site deployments.

After speaking with the App Volumes Product Manager, I am happy to confirm that, “Yes,” you can use this new feature to support multi-site deployments – as long as you are using the supported SQL database.

The architecture for this type of deployment would look like this:

DCarter App Volumes 2

 

I would recommend that App Volumes Managers at each site be clustered. Read the following blog to learn how to cluster App Volumes Managers: http://blogs.vmware.com/consulting/2015/02/vmware-appvolumes-f5.html

Although 2.9 is just a point release, this is one of the biggest features added so far for multi-vCenter support.

To add a second―or more―vCenter instance to App Volumes, follow these simple steps:

  1. Login to the App Volumes Manager
  2. Select Configuration, then Machine Manager, and then click Add Machine Manager
    DCarter App Volumes 3
  3. Enter the vCenter information and click Save.
    DCarter App Volumes 4
  4. Follow these steps for each vCenter instance you want to add.

Dale is a Senior Solutions Architect and member of the CTO Ambassadors. Dale focuses in the End User Compute space, where Dale has become a subject matter expert in a number of the VMware products. Dale has more than 20 years experience working in IT having started his career in Northern England before moving the Spain and finally the USA. Dale currently hold a number of certifications including VCP-DV, VCP-DT, VCAP-DTD and VCAP-DTA.

For updates you can follow Dale on twitter @vDelboy

Simple VDI Load Testing with View Planner

Jack McMichaelBy Jack McMichael, Solutions Consultant

In the last few years it seems the number of customers asking for assistance in re-evaluating their “VDI 1.0” infrastructure is increasing at a faster rate than ever. It makes sense when you consider that in the rush to achieve datacenter consolidation many administrators were under pressure to just “make it happen.” Many of those administrators and architects didn’t have time to design their virtual desktop infrastructure (VDI) solution to scale and accommodate things our customers and users have grown accustomed to using every day, such as YouTube HD, Skype, and other resource-intensive applications.

Last year, VMware released their very popular internal tool View Planner for use to the general public for free. While it’s flown under the radar for a lot of customers, it can be an invaluable tool for judging where your VDI solution stands, and identifying where the stress cracks in your VDI infrastructure may be forming—or are already wide open.

The View Planner appliance is simple to install and fairly straightforward to set up for local tests. It’s capable of local-only load testing, as well as passive/remote connections with the VMware Horizon® Client.

Deploying View Planner

After deploying an Open Virtual Appliance in VMware vSphere®, configure your View Planner integrations in the Config tab of the administrator page. The AD and View integrations are optional, but can be used if you wish View Planner to deploy desktops and/or create and delete users.

Note that for best results, I recommend using IP addresses instead of hostnames. Create a service account for your credentials, and give it administrator privileges in both AD and in VMware vCenter™.

In this screenshot, you can see all three connectors configured. You can use the Test buttons to ensure the configuration works, but click Save first.

JMcMichael 1

Environment Preparation

For my simple test, I created a linked clone pool with the name VPDesktop-{n:fixed=3} in VMware Horizon View™. On this master snapshot, I added the View Planner Desktop Agent that you can download from the View Planner portal on the Packages tab.

Make sure you reboot your desktop before creating your snapshot. Once you reboot, you will likely see the desktop auto-login. If so, run the View Planner Agent as seen in this screenshot.

JMcMichael 2

 

Configuring Run Profiles

There are three test modes available: Local, Passive and Remote. Typically, Local mode will be used for load testing since it doesn’t require actual Horizon Client connections, but has the disadvantage of not replicating PCoIP performance impact. Passive mode will add PCoIP connections that are shared amongst client servers that host more than one client connection at a time. Remote mode will create a 1:1 relationship between clients and desktops, thus creating the most overall resource impact.

To configure a Run Profile for a simple load test, I recommend using Local as it doesn’t require the use of the Horizon Client, and is also easy to set up. Simply add the Workload profile you want to run into the Run Profile by clicking Add Group, and click Save to save the Run Profile. You can add multiple workload profiles if you desire, but for a simple test only one is required.

The most important thing to remember is that desktop names (and client names if you choose Passive or Remote) are case-sensitive. In this example, VPDesktop– is valid for VPDesktop-001, but not vpdesktop-001 or VPDesktop001.

JMcMichael 3

Running a Test

Simply click the Run button to start a test. If you run into trouble, View Planner will show you right away; by clicking the link on the appropriate box, you’ll see the exact error or success message.

JMcMichael 4

 

Once completed, you can view the results in the Per Stats column; they will look something like the example below.

JMcMichael 5

Summary

Overall, I found the View Planner tool to be great for simple and quick tests of a VDI environment. It shows you where resource contention exists, or singles out how an app may be creating resource gaps in your VMware ESXi™ hosts. The free downloadable version includes several standard templates that cover a variety of normal user application workloads. If you require more flexibility in your tests, a paid VMware Professional Services engagement offers a more feature-rich version to create customizable workload profiles and other goodies. Contact VMware Professional Services or a VMware Partner for an on-site evaluation.

 


Jack McMichael is a Solutions Consultant for the VMware Professional Services Engineering Global Technical and Professional Services team. Follow him on Twitter @jackwmc4 !

So You Virtualized Your Desktop Environment. Now what?

mmarx.phpBy Mike Marx

Most of my customers start with a low-risk user group consisting of a large number of users with identical application requirements. This is the common scenario when starting out on the virtual desktop infrastructure (VDI) journey and ‘testing the waters.’ With proper design efforts, initial implementations are highly successful.

I spend the majority of my consulting effort working with customers helping them create their initial VDI design. Designs can be simple or complicated, but they all utilize a common technical approach for success: understanding user requirements, and calculating infrastructure sizing. But I’m not blogging about technical calculations or infrastructure sizing. Instead I would like to address a VDI design challenge customers face as they expand their VDI design: user application assignments.

While resource requirements are simple to assess, calculate and scale, application delivery becomes increasingly challenging as more users are added to the design. VDI administrators struggle to manage increasing numbers of desktop users – each having unique application requirements.

Applications are easy to add to a large static group of user desktops using linked-clones. But when unique user groups are introduced, and application requirements change, administrators are confronted with the challenge of maintaining a large number of small desktop pools – or impacting large groups of users in order to change an application assignment.

So how do we design an effective stateless desktop and maintain application diversity amongst unique user groups? VMware Horizon AppVolumes is the answer.

Using AppVolumes, VDI designs become simple to understand and implement. Once applications are effectively removed from the VDI desktop, VDI administrators are left with a simple stateless desktop. But users aren’t productive with an empty desktop operating system; they need applications – and lots of them.

Without going into deep technical detail (there are excellent blogs on this topic already) AppVolumes captures the application files, folders and registry components, and encapsulates them into a transportable virtual disk called an AppStack. As the user logs on to a stateless desktop, the assigned AppStack(s) will automatically attach and merge the user’s applications with the desktop virtual machine.

Now users are presented with a stateless desktop that is uniquely assembled with all of their applications. AppVolumes’ attached applications interact with other applications— and the operating system—as if they were natively installed, so the user experience is seamless.

Now that applications are no longer an impediment to VDI designs, VDI administrators are able to support large groups of users and application requirements using the same stateless desktop pool. By following the KISS principle: “Keep It Simply Stateless,” AppVolumes will open the door to new design possibilities and wider adoption by users and IT administrators.


Mike Marx is a Consulting Architect with the End User Computing group at VMware. He has been an active consultant using VMware technologies since 2005.  His certifications include : VCAP-DTD, VCP-DT, VCA-WM, VCA-DT, VCP2-5 as well as being an expert in VMware View, Thinapp, vSphere and SRM.

VMware User Environment Manager Deployed in 60 Minutes or Less

By Dale Carter

With the VMware acquisition of Immidio, announced in February 2015, VMware has now released VMware User Environment Manager (UEM). In the last several weeks I have been doing some internal testing with UEM and looking at the different things the software can do, and how this will help administrators manage users and improve the user experience.

After the acquisition was complete I kept hearing internal conversations about just how easy UEM is to deploy and get up-and-running, as there is no extra infrastructure needed to configure UEM. All that is required to configure UEM is:

  • A couple of file shares
  • Configuration of group policy objects (GPOs) on the User organizational unit (OU)
  • Installation of UEM agent and manager software

Unlike a lot of other management software, VMware UEM only requires that the software is installed on one or more administrator desktops. There is no management server component other than network file shares and configuration of a few GPOs.

Given the simplicity of the installation, I decided to document how easy it is to get an enterprise-ready solution deployed in less than 60 minutes. Now, this is a basic deployment for 50 linked clone virtual machines, but you’ll see just how easy it is to deploy and configure them. For an enterprise with many sites, decisions need to be made about configuring network shares and where to place them on the network. But most of the work, as you will see, can easily be accomplished in 60 minutes or less. Read the VMware User Environment Manager Guide below:

DCarter UEM 5

 


Dale is a Senior Solutions Architect and member of the CTO Ambassadors. Dale focuses in the End User Compute space, where Dale has become a subject matter expert in a number of the VMware products. Dale has more than 20 years experience working in IT having started his career in Northern England before moving the Spain and finally the USA. Dale currently hold a number of certifications including VCP-DV, VCP-DT, VCAP-DTD and VCAP-DTA.

For updates you can follow Dale on twitter @vDelboy

Link VMware Horizon Deployments Together with Cloud Pod Architecture

By Dale Carter

VMware has just made life easier for VMware Horizon administrators. With the release of VMware Horizon 6.1, VMware has added a popular feature—from the Horizon 6 release—to the web interface. Using Cloud Pod Architecture you can now link a number of Horizon deployments together to create a larger global pool – and these pools can span two different locations.

Cloud Pod Architecture in Horizon 6 was sometimes difficult to configure because you had to use a command line interface on the connection brokers. Now, with Horizon 6.1, you can configure and manage Cloud Pod Architecture via the Web Admin Portal, and this greatly improves the Cloud Pod Architecture feature.

When you deploy Cloud Pod Architecture with Horizon 6.1 you can:

  • Enable Horizon deployments across multiple data centers
  • Replicate new data layers across Horizon connection servers
  • Support a single namespace for end-users with a global URL
  • Assign and manage desktops and users with the Global Entitlement layer

The significant benefits you gain include:

  • The ability to scale Horizon deployments to multiple data centers with up to 10,000 sessions
  • Horizon deployment support for active/active and disaster recovery use cases
  • Support for geo-roaming users

This illustration shows how two Horizon deployments—one in Chicago and another in London—are linked together.

DCarter View 6.1

To configure Cloud Pod Architecture for supporting a global name space you first:

  • Set up at least two Horizon Connection Servers – one at each site; each server would have desktop pools
  • Test them to ensure they work properly, including assigning users (or test users) to the environments

Following this initial step you create global pools, then configure local pools with global pools, and finally set up user entitlements, which can be done from any Horizon Connection Server.

For more detailed information, and for a complete walk-through on setting up your Cloud Pod Architecture feature, read the white paper “Cloud Pod Architecture with VMware Horizon 6.1“.


Dale is a Senior Solutions Architect and member of the CTO Ambassadors. Dale focuses in the End User Compute space, where Dale has become a subject matter expert in a number of the VMware products. Dale has more than 20 years experience working in IT having started his career in Northern England before moving the Spain and finally the USA. Dale currently hold a number of certifications including VCP-DV, VCP-DT, VCAP-DTD and VCAP-DTA.

For updates you can follow Dale on twitter @vDelboy