With Hyper-V, it is very easy to over commit your storage space by creating dynamic VHD files. It is also easy to consume all of your available storage by creating multiple, parallel snapshots, creating large fixed VHD files, and letting your dynamic VHD files expand as the guest operating systems demand more space.

What happens when you hit the limit on your storage devices? Your virtual machines are placed into the dreaded “Paused Critical” or “Major Failure” states.

According to Virtual PC Guy :

If we were to run out of space for expanding a dynamic or differencing virtual hard disk – we would start failing disk write operations inside the virtual machine. The result would most likely be that the guest operating system would crash. To avoid this problem we periodically poll the storage that is being used for the virtual hard disks. If the free space falls under 200 megabyte we will pause the virtual machine and mark it as “Paused-Critical”.

This raises the unfortunate possibility of a mission-critical production application suddenly becoming unavailable because an unrelated virtual machine decided it needed more space.

Just what files do virtual machines need and where are they located? Again, Virtual PC Guy is a good source of information. In his articles “Understanding where your virtual machine files are in Hyper-V” and “Where are my snapshot files” he discusses the types of files virtual machines use, where Hyper-V stores them, and outlines some perils:

One thing that often trips people up is if they have their virtual hard disks configured on one drive – but have left their snapshot files stored on the system drive.

Savison’s Vital Signs for Hyper-V can help guard against virtual machines running out of storage space, can help you proactively manage your VHD placement, and can show you where you can free up space if you have, unfortunately, run out.

The KPI tab within our dashboard will allow you to monitor both physical storage space as well as disk errors reported by the virtual machines – a sign that their VHDs are not able to expand.

The storage tab graphically shows the occupancy of all drives broken down by virtual machine, shows all files used by your virtual machines, and sums up the total physical storage required by each virtual machine.

With this information in hand, you can make intelligent decisions about where to place new VHD files and to determine where you can free up space – whether by removing snapshots or moving virtual machines – for your mission-critical VMs.

For more information about this Blog article or Vital Signs contact us! You can read more blogs articles at blog.savision.com.

Nick Hebert, Account Manager, Savision: Why in a hospital setting is it extra important to have real-time monitoring information?

Errol E. Small, Technical Services Manager, HealthCare Partners Medical Group: In this type on environment near or real-time monitoring and notification could be the defining factor in a life or death situation between the Doctor and patient relationship, if information is not readily available. Having the dependency of accurate monitoring can ward off any issue before it becomes a critical incidence.

Rogee Fe de Leon, Technical Services Manager, Maimonides Medial Center: It is important because a downed system will impact the hospital work flows which in turn will affect Patient Care.

NH: Are you using Live Maps to monitor your critical hospital devices?

RFdL: Yes. Specifically, we use Live Maps to monitor the servers, routers, and switches which enable the use of the Hospital’s various applications.

ES: All critical production servers that support our environment are monitored by Live Maps

NH: How has Live Maps improved your support team’s access to SCOM data?

ES: Live Map places the tons of alerts and events that OpsMgr provides and make it easy to drill down to the root cause and provides explanations of cause and resolutions where possible.

RFdL: Live Maps allows us to see right away where a problem is and what particular SCOM monitor is reporting an issue.

NH: What is an extra feature Savision could add to Live Maps to make your lives easier?

RFdL: I would suggest adding more drawing features.  One feature I would suggest is the ability to draw tables.

ES: Expand your knowledge base and training videos

What recommendations do you have for Savision in their development of Live Maps for 2012 and Vital Signs? Click here to give your suggestion!

This situation only occurs when one or more object have (previously) been added to a Live Maps view that are defined in the Management Pack that you want to delete. When you follow the suggestion as presented in the dialog above, all Live Maps views that have been created will be deleted.

Obviously, in most situations this is not what we want as this would enforce us to recreate all our Live Maps views from scratch. In this article I will describe the process to follow to remove the dependency from the Live Maps Authoring Management Pack without losing all our other Live Maps views.

Where did that dependency came from?

When objects of a class as defined in another Management Pack are added to  a Live Maps view, either as explicit objects or through Membership Rules, a Reference to the defining Management Pack is added to the Live Maps Authoring Management Pack. In this example a Live Maps view is created which contains all SQL Server 2000 Databases by using a Membership Rule as follows:

When we now take a look at the XML export of the Live Maps Authoring Management Pack, we will find that specific references to the SQL Server 2000 Management Packs are added:

<Reference Alias="MicrosoftSQLServer2000Discovery6066480">
 <ID>Microsoft.SQLServer.2000.Discovery</ID>
 <Version>6.0.6648.0</Version>
 <PublicKeyToken>31bf3856ad364e35</PublicKeyToken>
 </Reference>
<Reference Alias="MicrosoftSQLServerLibrary614000">
 <ID>Microsoft.SQLServer.Library</ID>
 <Version>6.1.400.0</Version>
 <PublicKeyToken>31bf3856ad364e35</PublicKeyToken>
</Reference>

When we further explore the Live Maps Authoring Management Pack, we will find the Alias for these references used in the definition of the Membership Rule for the Live Maps group created above.

How to get rid of that dependency?

Before we can delete the SQL Server 2000 Discovery Management Pack, we need to remove all explicit objects and Membership Rules from all our Live Maps views through the Live Maps Authoring Console:

After removing them all, we can export the Live Maps Authoring Management Pack in the Administration pane of the Operations Console:

Make a copy of the unmodified exported XML file as a backup!

Then open the XML file in an XML Editor of your choice (e.g. Microsoft’s XML Notepad 2007) and make a note of the Alias used. Search the remainder of the XML file to assure that there are no Membership Rules that are using the alias.

Then delete the Reference section(s) to the Management Pack(s) that we want to delete:

After importing the modified Live Maps Authoring Management Pack into our Operations Manager environment again, the previously referenced SQL Server 2000 Management Pack can now be deleted:

If the import of the manually modified Live Maps Authoring Management Pack fails, it probably means that the Alias of a removed Reference is used anywhere else in the Live Maps Authoring Management Pack or that the XML structure itself became invalid. Please revert back to the previously created backup of the unmodified exported XML file and carefully repeat the steps described above.

More information on Live Maps can be found here.

To prevent sprawl, I would suggest one of the following two strategies. Ideally, both would be used.

The first would be to ensure that there are some checks in place for creating virtual machines.

· Does this virtual machine need to be monitored?

· Is the monitoring team aware of this new virtual machine and its monitoring requirements?

· Is this virtual machine being created to serve a temporary need?

· Who will be responsible for removing it later?

Answering some of these questions before creating the virtual machine can dramatically increase your chances of not losing control of it later.

The second approach would be to have a great tool for virtualization management in place. This tool should be able to assess the current capacity needs of your virtual infrastructure, as well as quickly identify stale virtual machines. Identifying stale virtual machines might not always be straight forward. But, if you have the right tools in place that can accurately present where resources are not being best utilized, you have a much better chance of keeping virtual machine sprawl under control. After all, the ability to be nimble with your virtual infrastructure is a large part of why it’s there in the first place.

Savision’s Vital Signs for Hyper-V can provide insight into the health and performance of your Hyper-V hosts, as well as which virtual machines are consuming the bulk of the resources. This information can be very useful in determining if your virtual machines are appropriately sized. Vital Signs will also graphically present the storage footprint used by virtual machines, as seen in the snippets below. This can streamline finding virtual machines that are consuming large amounts of storage, and help you identify opportunities to free up space in your storage subsystem.

Vital Signs for Hyper-V can also show all the virtual machines in your Hyper-V environment, and what applications are running on those virtual machines. Having that information in one view, greatly improves the ability to identify stale virtual machines that can be removed to free up resources.

Stay tuned over the next few months, as Savision will have some exciting announcements at MMS around how to better manage your Hyper-V environment.

If you want to contact us about Vital Signs for Hyper-V or something else, Click Here!

System Center Virtual Machine Manager 2012:
Cloud computing represents a new era of significant cost savings combined with unprecedented speed and agility. A core component of Microsoft’s cloud computing strategy is System Center Virtual Machine Manager 2012.

This new version of SCVMM presents a huge shift from managing virtual machines to managing the entire virtualized datacenter.

The shift is so dramatic that “System Center Data Center Manager” might be a more appropriate name. If you’re in IT, you need to know about all of the great new features in SCVMM 2012!

Attend this webinar to learn about the cool new features of SCVMM 2012, including:

- Infrastructure Enhancements
- Support for Microsoft Private Cloud
- Fabric Management
- Storage Management
- Provisioning

About The Speaker:
Bryan Dehner – Managing Consultant, Catapult Systems

Bryan is a Managing Consultant at Catapult Systems with over seventeen years experience in the computer technology field. He specializes in Microsoft technologies to provide solutions to businesses, primarily concentrating on IT Infrastructure solutions and specializing in virtualization technologies. He is a certified MCITP for Windows Server 2008 R2, Virtualization Administrator certification. Bryan recently attended the SCVMM 2012 Technical Airlift Program at the Microsoft campus in Redmond where he was able to work with the developers of the SCVMM 2012 beta software.

This is why we added a little feature to Live Maps v5.1.1 that will do some performance measuring and shows the results to the user.

In the following section I will explain how this feature can be used in the Live Maps web console and Windows Gadget.

Web Console

To enable the performance tracing add ‘?showtrace=true’ in the address bar right after the url.

After the Live Maps is started you see something like ‘Enabled Tracing on 2012-01-05T16:18:46.0061652Z [UTC] [1 (Startup) - 297 ms]’ above the view. The startup time is the time the application needed to initialize and load the views tree.

Drawing view

Now if you select a drawing view you see the following performance trace information:

[1 (Load MonitoringView object) - 146 ms]

[1 (Open view) {

  [2 (Check license) - 75 ms]

  [2 (Create menu urls) - 29 ms]

  [2 (Load unplaced members) - 33 ms]

  [2 (Generate drawing view) {

    [3 (Load drawing view data) - 128 ms]

    [3 (Render bitmap) - 20 ms]

    [3 (Disk I/O bitmap) - 17 ms]

    [3 (Generate hotspots) - 604 ms]

  } – 771 ms]

  [2 (Load alerts) - 26 ms]

} – 942 ms]

At first it looks kind of cryptic, but if you copy/paste the info to a text editor and reformat it a bit like above it is much more readable. As you can see some actions contain multiple sub actions that’s why every action starts with a number. 1 is a root action and a higher number is a sub action. If you sum all the times from the actions with number 1 you have the total load time. For this example it is 146 + 942 = 1088 ms.

Except for ‘Render bitmap’ and ‘Disk I/O bitmap’ all actions are retrieving data from Operations Manager. The reason why ‘Generate hotspots’ is taking a lot of time is that for all the objects on the view the tooltip text is created. And for this text all property values of that object are used.

If you wait some time the view will get refreshed and you get the following information:

[1 (Refresh) {

  [2 (Load unplaced members) - 74 ms]

  [2 (Generate drawing view) {

    [3 (Load drawing view data) - 131 ms]

    [3 (Render bitmap) - 20 ms]

    [3 (Disk I/O bitmap) - 17 ms]

    [3 (Generate hotspots) - 502 ms]

  } – 672 ms]

  [2 (Load alerts) - 20 ms]

} – 774 ms]

For the refresh a subset of actions, that are also needed for opening a view, is done.

List view

After a list view is selected you see the following information:

[1 (Load MonitoringView object) - 200 ms]

[1 (Open view) {

  [2 (Check license) - 110 ms]

  [2 (Create menu urls) - 71 ms]

  [2 (Load list members) - 1972 ms]

  [2 (Load alerts) - 235 ms]

} – 2390 ms]

In this example I have chosen for a slower network connection between the Live Maps web console and the RMS of Operations Manager. As you can see the time needed for loading data is about 140% higher. The total load time is much higher than the previews drawing view because there are a lot of Operations Manager objects on this view.

Windows Gadget

A particular customer had the feeling that the performance from a view opened in the Windows Gadget fly-out was slower than when it was opened in the normal Live Maps Web Console. That is why we made it possible to enable the performance trace functionality in the Windows Gadget fly-out.

This is how you can enable this feature in the Windows Gadget.

Open the settings.ini file from ‘%LOCALAPPDATA%\Microsoft\Windows Sidebar\Gadgets\livemaps.gadget\settings’

The content of the file looks something like:

Source code

Automatically generated by SettingsManager 
 
 vrijdag 6 januari 2012 9:48:25 
 
 [global]  server="http%3A//localhost%3A11424" 
 
 showTrace="true" 
 
 userName="Administrator" 
 
 domain="contoso" 
 
 integrated="false" 
 
 useBrowser="false"

Change the showTrace=”false” to showTrace=”true”.

If you now open the fly-out in the gadget you see the same performance trace information as described above for the Live Maps web console.

I hope this feature will help you get more understanding how Live Maps works in case you need this knowledge for solving performance issues.

This feature is available in Live Maps for System Center Operations Manager V5.1.1 and higher.

For more information about Live Maps click here, or contact us!

How do you install these tasks? Once you have carried out the one-time integration between Operations Manager and Vital Signs, simply bring up the context menu within Vital Signs and choose “Add Tasks”. Vital Signs will then insert the launch tasks into Operations Manager for you.

Are you running Live Maps to visualize your IT environment? If so, then you can launch Vital Signs from within your maps as well. Simply bring up the context menu, select “Open”, and then “Vital Signs.”

While this may be a minor feature, we believe it is important to do whatever we can to speed you on your way to problem resolution.

For more information about Vital Signs click here, or contact us!

Nick Hebert (Savision, head of Americas sales): What is the greatest value you see in the Live Maps product?

Seth Steward: “Bringing a visual component to a very data-driven side of technology.”

Ryan Durbin: “The largest value I see in Live Maps is the ability to add a tailor-made custom interface to SCOM with the ability to create as many custom views as you like that are easy to understand for non-techy individuals such as upper management, etc.”

NH: How has Live Maps improved employee access to critical SCOM data?

RD: “Live Maps has given me the ability to create an overview of the current state of health of our operations so that at any given time we feel confident in looking at the high level maps and knowing that we are doing well. When there is an issue we can then drill down into the maps as far as needed to find the issue and seek answers from our knowledgebase to engage the issue in a timely fashion.”

SS: “Live Maps has enabled our operations employees to be more efficient at responding to alerts because of the visual aspect Live Maps brings to SCOM. Instead of solely relying on e-mail alerts, our employees can now visually see which server in a rack is experiencing issues and with just a click, can see the alert and begin troubleshooting. Live Maps also provides context so you can easily see the trickle-down effect an outage on one or more devices may have.”

NH: Help me understand how Live Maps has changed your alert problem resolution flow.

RD: “Before Live Maps our operations department had a convoluted aspect of what the current health of our environments were. Now that we have Live Maps in place, the state is available at a glance.”

SS: “It’s much easier to be proactive once you have Live Maps for a system built out. Instead of being reactive to e-mail alerts, you can create dashboards of the systems you manage and glance at them frequently to ensure you’re in a healthy state. If you do receive an e-mail alert notifying you of an issue, Live Maps can help you quickly pinpoint the effect that issue may have as well as provide shortcuts for problem resolution such as reattaching a database or restart a web service.”

(Live Maps popular Drill-Down function)

NH: What are you most excited about for Live Maps 2012?!

RD: “The most exciting things for us as a company for 2012 would have to be Service Manager integration.”

SS: “Given our heavy focus and dependability on System Center products, the integration between Service Manager and Operations Manager in the next version of Live Maps is paramount to our business. The ability to actively sync our 500+ Business Services without having to wait for the SCOM CI Connector to run will ensure our CMDB is always up to date.”
(end of the interview)

Want to learn more about what Savision has in store for 2012? Click here to schedule your personal demo for Live Maps 2012, or download our Beta 2 Version Live Maps for SCOM 2012! Also learn more about Vital Signs, a product which is a realtime troubleshooting program and instantly pinpoints the root cause of your IT problems. Vital Signs also integrates seamlessly with Live Maps.

For more information contact us at Sales@savision.com!

To be able to also use Windows Authentication on a non-RMS server, we need to setup Kerberos constrained delegation. In this article the steps required to allow for Windows Authentication while having the Live Maps Web Console installed on a standalone Web Server are described.

Verify the Service Principal Name (SPN) Registration

A service principal name (SPN) is the name by which a client uniquely identifies an instance of a service. Kerberos authentication uses the SPN to authenticate the domain account used by the OpsMgr SDK Service, so first we need to verify the SPN’s attached  to the SDK account. In this example setup, the OpsMgr SDK Service (System Center Data Access) account used is: VSDEMO\Administrator.

We can check the SPN’s of this account as follows:

If, for any reason, these are missing, we need to add those. This article on Microsoft TechNet describes how to do this. In this setup it would require the following:

setspn /A MSOMSdkSvc/VSDEMO-OPSMGR VSDEMO\Administrator
setspn /A MSOMSdkSvc/VSDEMO-OPSMGR.VSDEMO.local VSDEMO\Administrator

Verify the Domain Functional Level

When setting up constrained delegation, the domain controller should operate at a Domain functional level equal to Windows Server 2003 or higher. This can be verified by launching the ‘Active Directory Domains and Trusts’ MMC snap-on and right-click the domain to select the Properties from the context menu:

Verify the SDK Service Account options

Verify the user account options by launching the ‘Active Directory Users and Computers’ MMC snap-on, select the Properties of the SDK Service Account and navigate to the ‘Account’ tab. Ensure that ‘Account is sensitive and cannot be delegated’ is NOT checked:

Configure the Web Server for Constrained Delegation

To configure the Web Server for Constrained Delegation, find the Computer Account of the Web Server where the Live Maps Web Console will be installed on. From the Properties context menu, navigate to the ‘Delegation’ tab and check the option ‘Trust this computer for delegation to any service (Kerberos only)’:

Install the Live Maps Web Console with Windows Authentication

As we have constrained delegation set up for our standalone Web Server, we can now install the Live Maps Web Console on this server and select ‘Windows Authentication’ during the installation:

If you previously installed the Live Maps Web Console using Forms Authentication on this server, it is advised to uninstall it using ‘Programs and Features’ and then reinstall it using ‘Windows Authentication’.

References

In this blog article from Kevin Holman a similar procedure is described for the installation of the OpsMgr Web Console with Windows Authentication on a non-RMS Server. This blog article also lists some details on achieving the same on a Windows 2003 server.

More details on Live Maps for System Center Operations Manager can be found here.

• Task Manager on the Hyper-V host
• Task Manager on each virtual machine (assuming they are running Windows)
• Hyper-V Manager
• Performance Monitor through the Processor, Hyper-V Hypervisor Logical Processor, Hyper-V Hypervisor Root Virtual Processor, and Hyper-V Hypervisor Virtual Processor object

It can be difficult to grasp what each of these mean, and ultimately determine if the processors on your Hyper-V host are saturated.  Let’s go through each one in more detail.

1) Task Manager on the Hyper-V Host

Outside of Hyper-V, determining CPU usage on a Windows Server is pretty straight forward when using Task Manager.  However, this all changes if the Windows Server is also a Hyper-V Server.  I’ve seen many users of Hyper-V get confused when they open Task Manager on a busy Hyper-V host to see a very low number reported as the current CPU usage.  This Hyper-V host runs many busy virtual machines, so how can this be?  The key here is that when the Hyper-V Server role is installed, the Windows OS that runs Hyper-V is actually just another virtual machine (albeit a special virtual machine).  Since a virtual machine has no idea what resources are being consumed by other virtual machines, what you see in Task Manager on the Hyper-V host only pertains to the CPU being used by this special virtual machine, and not to the physical server or any of the other virtual machines being managed by Hyper-V.  In general, this method is not terribly useful in determining how busy your Hyper-V host is.

2) Task Manager on a Virtual Machine

Looking at Task Manager on a virtual machine will tell you how much CPU is being used by that virtual machine.  However, what does this really mean?  A virtual machine is allocated some number of virtual processors, and these virtual processors will share the Hyper-V host’s logical processors with the other running virtual machines.  Task Manager in this case, will let you know how much of the virtual processors are being used, but provides no insight into how this is affecting the logical processors on the Hyper-V host.

3) Hyper-V Manager

It’s easy to get confused when looking at the CPU Usage as reported in Hyper-V Manager.  As an example, consider a virtual machine that shows 100% CPU in Task Manager from within the virtual machine.  However, in Hyper-V Manager, it only shows 11-12%.  What is the reason for the large discrepancy?  Hyper-V Manager is showing CPU Usage of the virtual machine, but as a fraction of the available logical processors on the Hyper-V host.  So, when a virtual machine that has 1 virtual processor is pinned at 100% on a Hyper-V host that has 8 logical processors, Hyper-V Manager will show CPU Usage as roughly 12% (100% / (8 logical processors / 1 assigned virtual processor)).  If one were to assign 2 virtual processors to this same virtual machine, and the CPU would remain pinned at 100%, Hyper-V Manager would now show roughly 25% (100% / (8 logical processors / 2 assigned virtual processors)).  Unless you have memorized the number of logical processors on all your Hyper-V hosts, and the number of virtual processors assigned to all of your virtual machines, quickly making sense of this data is challenging to say the least.

4) Performance Monitor

As mentioned above, there are a number of performance objects that report CPU usage.  I’m going to focus on the Hyper-V Hypervisor Logical Processor and Hyper-V Hypervisor Virtual Processor objects.  The Hyper-V Hypervisor Logical Processor object shows how much of the logical processors are being used, and the Hyper-V Hypervisor Virtual Processor object shows how much of the virtual processors are being used.  Within each of these performance objects, the % Guest Run time counter is the most important.  It shows how much of the logical or virtual processor is being used by virtual machines.  Looking at this counter for the Hyper-V Hypervisor Logical Processor object will give you an accurate picture of how much of the logical processors are being used to serve the needs of the virtual machines.  Tracking this counter is key to ensuring optimal performance of your virtual machines.

In the context of the previous example, the Hyper-V Hypervisor Virtual Processor – % Guest Run time counter is reporting values of 90+%, which is a much more accurate representation of what is actually occurring within the virtual machine.  The Hyper-V Hypervisor Logical Processor – % Guest Run time counter is reporting roughly 25%, providing insight into how much of the logical processors are being used to serve the needs of the running virtual machines.

Of course, instead of trying to remember which counters are important and what they mean, you could use Savision’s Vital Signs for Hyper-V Server to do that for you!  Vital Signs does the heavy lifting of aggregating the right counters for you, so you can quickly determine how busy your Hyper-V host is, and what virtual machines are contributing to that load.

With Vital Signs, it is easy to see the impact virtual machines are having on a Hyper-V host as shown below.

Additionally, moving to the Virtual Machines page within Vital Signs, one can also easily determine which virtual machines are responsible for the bulk of the load on a Hyper-V host.

You can evaluate Vital Signs in our online demo environment right now, by visiting http://www.savision.com/login-vitalsigns.