Latest Blogs

Chargeback Reporting for MS Private Cloud with Operations Manager 2012

Pete Zerger — Wed, 30 May 2012 14:36:58 GMT

This is a largely ignored topic that is fundamental to MS private cloud operation and management I think deserves some attention.

Since the Cloud Services Process Pack is not a solution that fits the bill for all (or even most) MS private cloud deployments, the need arises for an alternate method of calculating resource consumption for business units across one or more clouds as defined in System Center Virtual Machine Manager 2012 (VMM). As part of the ‘private cloud in the real world’ series, I am going to show you a few reports I leverage to answer the difficult questions around resource consumption and workload distribution in a private cloud deployment.

The most easily accessible reports actually come in System Center 2012 Operations Manager (OpsMgr) by way of the System Center 2012 Virtual Machine Manager (VMM) Management Pack for OpsMgr, which contains a number of reports for tracking the logical and physical consumption of resources in your private cloud fabric (the storage, network and compute). When leveraged appropriately, a good source of showback information for the business units subscribing to resources in your private cloud infrastructure. And with native SharePoint integration and flexible delivery options, they can become an easily accessible source of information for stakeholders in your cloud environment who need information, but are less familiar with the System Center tools.

The reports available in the VMM 2012 Management Pack for OpsMgr are pictured in figure 1. In this installment, we’ll take a look at the Chargeback report.

figure 1. Available reports in the VMM 2012 Management Pack

Report Parameters

The parameter header in the Chargeback report provides a number of fields that must be completed to provide meaningful results, including:

Date Range:The date picker is straightforward…I like ‘First day of previous month’ to ‘ last day of previous month’ as a nice rolling dataset of resource usage in the cloud.

Objects: This is where you can add your virtual machines that are (or were in the past) deployed to the cloud.

Billing Units: You can assign a cost to Memory, CPU, Base Cost for VM. It's easy to making things too complicated with this setting. I suggest that, at least in the first iteration, you pick one of these units is your billing mechanism for chargeback in order to keep the results meaningful. You can add multiple billing units as your organization progresses in their use the chargeback.

figure 2. Smart parameter header in the Chargeback report

Tip: I'd suggest creating a linked report (which will require some XML) in order to make this a dynamically populated report.

The result of a sample reporter shown below, which logically groups are infrastructure based on 1) Cloud and 2) Host Group, both of which are logical constructs defined in VMM 2012.

figure 2. Report content in the Chargeback report

Even if you don’t make this a dynamically populated report,You can say this report your favorites, making it a one click report with the occasional tweak for new infrastructure edit your cloud.

Additional Resources

Hopefully this info provides some food for thought when choosing your naming conventions for your template-based deployments in VMM 2012. Here are a few additional articles in the ‘private cloud in the real world’ series and related System Center articles you may find interesting.

Creating and Using Custom Properties in VMM 2012 in PowerShell (and a gotcha)

Pete Zerger — Tue, 29 May 2012 06:45:26 GMT

Custom properties, introduced in Virtual Machine Manager 2012 (VMM), consist of a user-defined name / value pair. They can be used in a variety of circumstances including

VM placement (or use of matching properties on host groups),
Reporting (by using custom properties to tag VM's and services with various bits of owner information)
Lifecycle Management (through the addition of a custom property to store expiration date)
Customizing View in the VMM Admin Console (more on this in a minute)

While you can add custom properties in the UI, I like adding them to PowerShell, which I can incorporate into Orchestrator runbooks to automate the VM and service provisioning, tagging by virtualization assets to support all of the possibilities mentioned above. Below is a sample script demonstrating how to tag a service instance with information to support virtualization lifecycle management, specifically owner information (a RequestedBy property) and an ExpirationDate property to store the retirement date for the newly deployed asset.

Sample Script

PowerShell snow below intended for use in a VMM script, presumably as part of an Orchestrator runbook. The only thing you need to do make the sample work for you is define is the value of a variable called $SvcName, intended to represent the display name of the newly deployed service (from a service template)

#-------------------------------------------------------------

#Set Custom Properties on Service Instance

#-------------------------------------------------------------

$MySvc = get-scservice | where {$_.Name -like $SvcName}

$customPropRequestedBy = Get-SCCustomProperty -Name "RequestedBy"

Set-SCCustomProperty -AddMember @("ServiceInstance") -CustomProperty $customPropRequestedBy

Set-SCCustomPropertyValue -InputObject $MySvc -CustomProperty $customPropRequestedBy -Value "$RequestedBy"

$customPropExpiry = Get-SCCustomProperty -Name "ExpirationDate"

Set-SCCustomProperty -AddMember @("ServiceInstance") -CustomProperty $customPropExpiry

Set-SCCustomPropertyValue -InputObject $MySvc -CustomProperty $customPropExpiry -Value "$VMExpiration"

Exposing Custom Properties in the UI

As alluded to earlier , in the VMs and Services pane of the VMM 2012 Admin Console, custom properties can be exposed by right-clicking on the title bar and selecting the appropriate properties. below is a screenshot the results of the sample PowerShell snippet shown above use as part of a runbook to deploy a service with VMM 2012.

Something to be aware of (Gotcha)

Custom attributes do not accept special characters, such as apostrophes, which I witnessed firsthand recently. There is a bug in VMM 2012 that presents itself when certain special characters are used in custom attributes throughout the UI, causing the console to crash. The fix should you encounter this problem is to remove the offending value via PowerShell or direct database edit, (through Microsoft support would recommend the PowerShell method).

Additional Resources

Tuning the System Center 2012 Virtual Machine Manager management pack

Cameron Fuller — Sat, 26 May 2012 20:54:21 GMT

Based upon my experiences with the SCVMM management pack, I added several ReSearch This! KB’s on this topic which are linked below:

Tuning for the System Center 2012 Virtual Machine Manager management pack: (alert name, URL for the ReSearch This KB)

VGS Not installed: http://www.systemcentercentral.com/BlogDetails/tabid/143/indexId/93620/Default.aspx

Host has exceeded the supported maximum number of running VMs: http://www.systemcentercentral.com/BlogDetails/tabid/143/indexId/93621/Default.aspx

Host Memory Utilization is high: http://www.systemcentercentral.com/BlogDetails/tabid/143/indexId/93622/Default.aspx

OM Agent Proxy not enabled: http://www.systemcentercentral.com/BlogDetails/tabid/143/indexId/93623/Default.aspx

ReSearch This! KB - VGS Not installed

Cameron Fuller — Sat, 26 May 2012 20:35:57 GMT

Alert: VGS Not installed

Management Pack Name: System Center 2012 Virtual Machine Manager Monitoring

Management Pack Version: 3.0.6005.0 and earlier

Issue: This alert occurs on every virtual machine in the VMM environment including VM’s which are offline. It looks like the Error condition detection will always be ‘true’ because this property (VGS) is an IC version and will never be ‘Detected’.

Resolution: Bug in the management pack. Disable this monitor until either a CU or a newer version of the management pack fixes this issue, storing the override in the appropriate override management pack (Virtual Machine Manager Overrides as an example). After disabling this monitor, the alerts should close themselves automatically.

Cloud: Computer Name Scavenging, Reuse and Conflict Handling in VMM 2012

Pete Zerger — Sat, 26 May 2012 03:33:21 GMT

Here are some notes from the field as part of our “Private Cloud in the real world” series. In this installment, I’ll talk about System Center Virtual Machine Manager 2012 (VMM) behavior when issuing names to VMs, reuse of names, conflict resolution (when no names are available) and the domain join process. I thought this useful info for those scenarios you may never have time to test for yourself, but for which the answers can prove very useful when known in advance.

Defining Available Name Values

When you design templates and provide naming conventions for VMs in VMM 20112, you can use wildcards (specifically, the # sign) in your templates as an indicator to VMM which characters it can assign unique numeric sequences to at deploy time. For example, a value of WEBAPP### would result in computer (NetBIOS) names of WEBAPP001 through WEBAPP999 being issued to to newly deployed VMs. This creates a range (or pool) of names that can be used (and re-used) in deployment.

VM Name Scavenging and Reuse

Default behavior (witnessed in deployment and deletion of more than 200 virtual machines this week) indicates that VMM scavenges names (returns names to the pool) when VMs are deleted and reuses them as soon as possible. For example, if you deploy 10 machines with the WEBAPP### convention using a service template (containing a single VM), they will be named WEBAPP001 – WEBAPP010. If you delete the service containing WEBAPP005, the next VM deployed will be named WEBAPP005.

This is a clear indicator that VMM is paying attention and making efficient use of available names based on the naming pattern provided.

VMM Behavior in Naming Conflicts

There are a couple of scenarios to consider here. Conflicting names and also poor choice in naming pattern that results in a limited pool of available names, in which case we may actually run out of available names for new VMs.

First, when you attempt to deploy from template and provide a name that is already in use, VMM will throw an error and block the attempt, which is predictable I think.

For the second scenario, let’s look at an example. I define a slightly different naming pattern in my template, this time WEBAPP#. This provides a pool of exactly nine names…WEBAPP1 – WEBAPP9. So what happens when the pool of available names is exhausted and you attempt to deploy a service (via service template) that contains the 10th machine?

VMM will actually launch the service deployment, but when the deployment attempts to provision the VM, the following error is thrown and the deployment fails.

Error (630). The computer name must be a valid IP address or name containing letters and numbers. The computer name cannot contain spaces or any of the following characters….

The takeaway here is simply that you want to consider how many VMs may be deployed simultaneously in your environment and choose a pool larger than you will need. Using the example above:

WEBAPP## gives you 99 unique options (WEBAPP01 – WEBAPP99)
WEBAPP## gives you 999 unique options (WEBAPP001 – WEBAPP999)

Domain Join with Name Reuse

My VM and service templates are configured to join the computer to the domain at deployment time. In a Windows 2008 R2 Active Directory domain, I found the domain join process was seamless, even with frequent name reuse. DNS could be an issue if security is tight (and not just anyone can update an A record), so I am running with a less stringent DNS security configuration in my environment. The most common issue I have encountered in conditions of frequent name reuse (not uncommon in a lab environment where lots of deployment testing goes on) is actually DNS cache on the Hyper-V host. I have seen connection to the VM after deployment fail, causing application deployment to fail, requiring the job to be restarted. Not an everyday occurrence, but something to consider.

Additional Resources

Orchestrator Jumpstart Series

System Center 2012–using a common SQL backend database (#SYSCTR, #SCOM, #SCSM)

Cameron Fuller — Fri, 25 May 2012 18:05:36 GMT

Now that System Center 2012 is one “product” we’re seeing a heavier focus towards using a common back-end for the products. Microsoft has done a good (not great) job on providing a consistent set of back-end database requirements so that a single version of SQL 2008 R2 SP1 CU6 should be sufficient for all of the System Center components.

<Start disclaimer here>

Before really getting into this let me state up front – this is not the approach that every environment will want to take. Some environments will require more dedicated resources including breaking down separate System Center components onto different database servers. Other environment (such as a lab) may be small enough that all of the databases could be configured to run in a single instance (if reporting services conflicts between the components was not a constraint).

<End disclaimer here>

In terms of database sizing, the largest databases we are seeing in our labs (and those requiring at least the most disk resources) are Service Manager, Operations Manager and Configuration Manager. Based upon lab testing up to this point, the following is the breakdown of database size for a lab environment which includes 25 servers, no workstations and a small number of users.

System Center Service Manager (ServiceMgr): 25 GB
System Center Operations Manager (OpsMgr): 12 GB
System Center Configuration Manager (ConfigMgr): 6 GB
System Center Data Protection Manager (DPM): 1 GB
System Center App Controller (App Controller): < .1 GB
System Center Virtual Machine Manager (VMM): .2 GB
System Center Orchestrator (Orchestrator): .1 GB

What about IOPS? (Input/output Operations Per Second)There is a lot of discussion on the number of IOPS required for various System Center components, do they really require the number of IOPS specified in the sizer? Sometimes a picture is worth a thousand words. The next picture shows the IOPS for a database server with about 2000 agents that has LUNS shared for the OperationsManager database and the OperationsManagerDW.

So… That 1000 IOPS estimation (mentioned later in this article) on the data warehouse for OpsMgr doesn’t seem so far fetched now! (total hovering around 1000, primarily write but with a significant amount of read)

The following is IOPS for the same environment (approximately 2000 agents) just for the OperationsManager database: (total hovering around 400, primarily write)

And the IOPS for the same environment for just the OperationsManagerDW: (total hovering around 200, primarily write)

What about SQL Collation? A blog post was recently added on this topic which is available at:

http://blogs.technet.com/b/servicemanager/archive/2012/05/24/clarification-on-sql-server-collation-requirements-for-system-center-2012.aspx

Can System Center components share an instance? There is not documented requirements to have a separate instance for each component of System Center. As long as the requirements for each database are met they can share but you need to take scale and performance into consideration. SRS is the one exception to this rule that I am aware of, requiring separate instances of reporting services for each component of System Center that uses SRS.

Case Studies:

For this blog article we are showing two similar case studies.

Case Study #1:

As a first case study example we installed a single two-node cluster to provide System Center back-end database functionality for ConfigMgr, OpsMgr, ServiceMgr and DPM. The client has approximately 200 servers that they are monitoring, and they have approximately 3000 workstations that they are managing and there are 5000 users in the environment. They are interested in potentially deploying all of the System Center components off of this back-end database. The following are the key pieces of the proposed back-end database:

Installing two physical servers on Windows 2008 R2 which are running a quad core with 32 gb of memory.

The two servers are connected to a SAN infrastructure to store all of the database and log files.

Instances are created on a per-component basis:

ConfigMgr
OpsMgr
ServiceMgr (2x)
DPM

Each instance is configured with separate spindles on the SAN for that instance’s database and log files.

Creating a separate reporting instance for each component that requires it (OpsMgr, ConfigMgr, Service Manager)

This cluster will only be for System Center back-end functionality.

Disk Sizing for Case Study #1: (based upon the Microsoft sizers)

ServiceMgr: 225 GB
OpsMgr: 160 GB
ConfigMgr: 100 GB
Total: 341 GB

Case Study #2:

As a second case study example, speaking with a client recently, we came to the consensus of creating a single two-node cluster to provide all System Center back-end database functionality. The client has approximately 2500 servers that they are monitoring, and they have approximately 6500 workstations that they are managing and there are 3500 users in the environment. They are interested in potentially deploying all of the System Center components off of this back-end database. The following are the key pieces of the proposed back-end database:

Installing two physical servers on Windows 2008 R2 which are running two socket/six core with 96 gb of memory.
The two servers are connected to a SAN infrastructure to store all of the database and log files.

Instances are created on a per-component basis: (gathering smaller sized databases into a single instance)

ConfigMgr
OpsMgr
ServiceMgr (2x)
DPM, App Controller, VMM, Orchestrator

Each instance is configured with separate spindles on the SAN for that instance’s database and log files.

Creating a separate reporting instance for each component that requires it (OpsMgr, ConfigMgr, Service Manager)

This cluster will only be for System Center back-end functionality.

Disk Sizing for Case Study #2: (based upon the Microsoft sizers)

ServiceMgr: 225 GB
OpsMgr: 1900 GB [Based on lab tests performed up to this point in time, it is reasonable to believe that the estimate from the sizer may be over-estimating for OpsMgr. We have one client with 2000 servers and a fully populated data warehouse in OpsMgr 2007 R2 has not crossed 500 GB yet, and another with about 1100 servers that has not crossed 1000 GB yet]
ConfigMgr: 100 GB
Total: 2225 GB

Sizing per Microsoft estimation methods for case study #2: (details on ServiceMgr, OpsMgr and ConfigMgr are contained in the appendix of this blog article)

ServiceMgr: 225 GB (http://www.microsoft.com/en-us/download/details.aspx?id=27850 or http://t.co/8ht9rAOQ)

OpsMgr: 1900 GB (http://www.microsoft.com/en-us/download/details.aspx?id=29270)

ConfigMgr: 100 GB (http://technet.microsoft.com/en-us/library/hh846235.aspx#BKMK_ReqDiskSpace)

DPM: .9 GB (http://technet.microsoft.com/en-us/library/hh757757.aspx)

App Controller: 1 GB (http://technet.microsoft.com/en-us/library/gg696060.aspx)

VMM: 150 GB (http://technet.microsoft.com/en-us/library/gg610574.aspx)

Orchestrator: 1 GB (http://blogs.technet.com/b/meamcs/archive/2011/11/02/orchestrator-2012-deployment-amp-configuration-step-by-step-part-2-system-requirements.aspx)

In terms of database sizing from the Microsoft sizing estimates, the largest databases (and those requiring the most resources) are Operations Manager, Service Manager, Virtual Machine Manager and Configuration Manager which represent the statistically relevant database sizes.

Common System Center 2012 Database Recommendations:

Limit each SQL Instance memory where the SQL total memory setting allows a minimum of 1 GB and preferably 2 GB of free memory after all services are running.
SQL Server will need to be Standard or Enterprise edition (Datacenter in some situations)
Install dual gigabit network adapters teamed with full duplex
Disks should be formatted to a 64K allocation unit size
Disks should be configured to allow 20% free space after all calculations for monitoring and set growth upper limits to not exceed this space limit.
The SQL Cluster should be running Active-Passive unless resources are proved to be insufficient
Common objections to sharing databases is contention for resources resulting in bottlenecks – processors, memory, and especially disk spindles. Use Operations Manager to identify these potential bottlenecks.
OpsMgr will not support SQL dynamic ports which is the default SQL setting, change each SQL instance dynamic port to a TCP port.
If OpsMgr will be providing monitoring for network devices, increase the size of the SQL temp in that instance.
Set the SQL Database Engine Domain Account(s) to auto register the SQL SPN values which occurs every time the service starts up
If you are using heart beat and iSCSI connections – make sure to remove all connection options except IPv4 and remove any DNS and Gateway entries as well.

Summary: Do you want to use a common SQL server for multiple components of System Center 2012? You can, but be sure to consider the requirements of each component and see the guidance above for sizing, instance details and general recommendations.

Special thanks on this one to: Travis Wright from the Microsoft product team for his second perspective, Louis Oliver for his help with the IOPS statistics, Chris Ross for his assistance with this from a Service Manager perspective!

Appendix: Sizing for OpsMgr, Service Manager and ConfigMgr from the Microsoft sizing estimators.

OpsMgr 2012 sizing:

In terms of currently available tools, there is a sizer for OpsMgr available at: http://www.microsoft.com/en-us/download/details.aspx?id=29270.

For the OpsMgr database in an organization this size the results are:

(Bolded Red items below are showing the Case Study #2 example with approximately 2000 OpsMgr agents)

DB Estimated Random IO Per Second for Maximum Load Configuration [80% Write, 20% Read]
# of Agents	Estimated IOPS
1-500	250
501-1000	500
1001-3000	750
3001-6000	1125
6001-10000	1250
10001-15000	1500

OpsMgr DB: Number of Days for Data Retention	7
Number of Server Computers	2500
Number of Network Devices	50
Number of APM-enabled Computers	0
Total Size (MB)	42276.18
Total Size (GB)	41.29
Total Size (GB) with 50% Buffer	61.93

For the OpsMgr DW in an organization this size the results are:

DW Estimated Random IO Per Second for Maximum Load Configuration [80% Write, 20% Read]
# of Agents	Estimated IOPS
1-500	500
501-1000	875
1001-3000	1000
3001-6000	1500
6001-10000	2000
10001-15000	2500

OpsMgr DW: Number of Days for Data Retention	365
Number of Server Computers	2500
Number of Network Devices	50
Number of APM-enabled Computers	0
Total Size (MB)	1707732.56
Total Size (GB)	1667.71
Total Size (GB) with 10% Buffer	1834.48

ServiceMgr 2012 sizing:

http://www.microsoft.com/en-us/download/details.aspx?id=27850

Service Manager database Size
Retention for Work Items (days)			90
Number of work items for computers/month	6150
Total Size (GB)			6.84
Suggested Space Allocation
with 50% buffer (GB)			10.26

Service Manager data warehouse Size Note: The details below for each database include the recommended 50% buffer.
			We also recommend placing each database on a separate physical drive if possible to optimize performance.
Number of Days			1095
DWStagingandConfig			20.53
Repository			61.58
DWDataMart			61.58
TempDB			51.32
DWASDataBase			18.47
		Note: The DWASDataBase size approximation is based on the default aggregation level set by Service Manager.

Total Size (GB)			213.49

Suggested Space Allocation			213.49
with 50% buffer (GB)

From download.microsoft.com/.../SC2012_ServiceMgr_Planning.docx

Service Manager database

Dual Quad-Core 2.66 gigahertz (GHz) CPU

8 gigabytes (GB) of RAM for 20,000 users, 32 GB of RAM for 50,000 users (See the Hardware Performance section in this guide.)

80 GB of available disk space

RAID Level 1 or Level 10 drive*

Data warehouse databases

Dual Quad-core 2.66 GHz CPU

8 GB of RAM for 20,000 users, 32 GB of RAM for 50,000 users (See the Hardware Performance section in this guide.)

400 GB of available disk space

ConfigMgr 2012 sizing:

from http://technet.microsoft.com/en-us/library/hh846235.aspx#BKMK_ReqDiskSpace

Data usage	Minimum disk space1	25,000 clients	50,000 clients	100,000 clients
Operating system	See guidance for the operating system.	See guidance for the operating system.	See guidance for the operating system.	See guidance for the operating system.
Configuration Manager Application and Log Files	10 GB	10 GB	10 GB	10 GB
Site database .mdf file	75 GB for every 25,000 clients	75 GB	150 GB	300 GB
Site database .ldf file	25 GB for every 25,000 clients	25 GB	50 GB	100 GB
Temp database files (.mdf and .ldf)	As needed	As needed	As needed	As needed
Content (distribution point shares)	As needed	As needed	As needed	As needed

How to convert a VMware VM Guest to Hyper-V in VMM 2012 (without vCenter integration)

Pete Zerger — Thu, 17 May 2012 13:47:52 GMT

Problem: I had to convert a VMware VM guest in my VMM 2012 test environment, but the virtualization team would not let me connect to vCenter from System Center 2012 Virtual Machine Manager. This means a virtual-to-virtual (V2V) conversion is not possible.

Solution: As a workaround, I decided to treat the VMware guest as a physical machine and perform a physical-to-physical (P2V) conversion.

FAQs of P2V Conversion in VMM 2012 <<< (good explanation of differences in ‘offline and online conversion)

P2V Wizard Selections

1 – In the VMs and Services workspace, Open VMM 2012 and go to home tab, then click on Convert Physical Machine under Create Virtual Machine button on ribbon.

2 - Put Server’s IP or name, followed by domain and login credentials. Keep in mind you need administrative rights on physical server (the VMware VM in this case) and it needs to be accessible from VMM server.

3 - Click Next and provide the name of the VM as it will appear in the VMM console.

4 - Click on Scan, which will then take a few minutes to scan the source VMware VM guest. Clicking on Scan installs an VMM agent on physical machine to gather the details about hardware and software on the source.

5 - Click next to proceed after finishing a scanning process. Now choose partitions you want to convert. You can also choose disk type (e.g. dynamic or fixed).

Note: System reserved volume is required, otherwise your new VM will not able to boot properly - this volume is preselected and greyed out to prevent mistakes. You will find you cannot set a smaller volume size (regardless of free space)…only larger

6 – Select number of processors and amount of RAM you want to assign the VM. Click Next to proceed.

7 – Intelligent placement in VMM will present suitable and available host in descending order based on available resources. Choose one and proceed by clicking on next to place you newly converted VM.

8 – Next, choose storage location for VM.

9 – Select the network interface to connect the VM to the same network as in the VMware environment. Click Next to proceed.

10 - Next choose automatic actions. In this VMM will validate all steps and it will display issues if any blockers are found. You must resolve any issues identified before clicking Next. If no issues are found, you are ready to proceed with P2V process.

11 - Make sure everything is correct on the Summary page. Do not choose to start automatically after deploy to avoid confliction between the old physical server and newly converted virtual machine. Click on create button to start p2v process. It will take some time according to your network speed and size of the source VMware guest. After clicking the Create button you can see the action in process under Jobs, where you can monitor progress of the conversion.

12 – When the P2V process is complete, turn off the source VMware VM guest before starting the Hyper-V guest to avoid network conflicts (since the newly converted machine is a clone of physical server).

Additional Information on System Center 2012 and MS Private Cloud

Here are a few of the many System Center and MS Private Cloud-related System Center 2012 articles here on SCC

Implementing rapid scale-out of a machine tier in VMM 2012 via PowerShell

Cloud: A quick note on accurately measuring Hyper-V host performance and utilization

What’s new in the Cloud Services Process Pack (CSPP) Release Candidate for System Center 2012?

Private Cloud in the real world: 5 lessons from the private cloud fabric

How to prepare for Orchestrator? Learn Opalis in a month of Friday afternoons! (free online resources)

PowerShell Scripts for System Center (Master Collection)

[Private Cloud]: Configuring anti-affinity between VMs on Hyper-V 2008 R2 SP1 and VMM 2012

SCVMM Tip: Testing PRO Integration in VMM / OpsMgr 2012

4 Hyper-V performance testing tips for the private cloud

Hyper-V Network Optimizations and Best Practices [Private Cloud]

Streamlining Linux VM configuration on Hyper-V with the Set-LinuxVM PowerShell module

KB2686249: Update Rollup 1 for System Center 2012 (KBs and hotfixes)

Tommy Gunn — Wed, 16 May 2012 15:33:09 GMT

This update rollup contains updates for the entire System Center 2012 family:

Virtual Machine Manager Server Update (KB2663959)
Virtual Machine Manager Console Update (KB2663960)
Virtual Machine Manager Self-Service Portal Update (KB2663961)
App Controller Update (KB2677596)
Operations Manager Update (KB2674695) – Windows monitoring
Operations Manager Update (KB2674695) – Unix and Linux monitoring
Operations Manager Update (KB2674695) – New feature

See http://support.microsoft.com/kb/2686249

Download

Here at the download links per-product

Virtual Machine Manager Server Update (KB2663959)

http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2663959

Virtual Machine Manager Console Update (KB2663960)

http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2663960

Virtual Machine Manager Self-Service Portal Update (KB2663961)

http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2663961

App Controller Update (KB2677596)

http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2677596

Operations Manager Update

http://www.microsoft.com/downloads/details.aspx?FamilyID=c593536c-94f3-4c71-b963-5ab3b7fc05e1

Hyper-V on Windows 2012 New Features and Enhancements [Video Shorts]

Pete Zerger — Tue, 08 May 2012 16:06:05 GMT

Here is a series of video shorts from some of the MVPs on the Infront team demonstrating some of the new virtualization-related features in Hyper-V v3 on Windows Server 2012 (aka Windows 8) including: Hyper-V Replica, Storage Live Migration, Controlling Network Bandwidth for VMs, Dynamic Memory Enhancements and Failover-related topics.

To access:

Click the link
Scroll to the top of the page (Youtube renders kind of funky in IE requiring this sometimes)

The Series

Configuring and Operating Hyper-V Replica in Windows Server 2012

Initiating a Failover in Windows Server 2012

Reversing the Replication Path after Failover in Windows 2012 Hyper-V

Dynamic Memory Enhancements in Windows Server 2012

Controlling Network Bandwidth for VMs in Windows Server 2012

Storage Live Migration in Windows 2012 Hyper-V

System Center 2012 Update Rollup 1 - SCVMM & App Controller

Tommy Gunn — Wed, 02 May 2012 04:33:13 GMT

There are more than 15 fixes in this first rollup. The downloads are available on Microsoft Update and at the following locations:

Virtual Machine Manager Server Update (KB2663959) - http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2663959
Virtual Machine Manager Console Update (KB2663960)- http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2663960
Virtual Machine Manager Self-Service Portal Update (KB2663961) - http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2663961
App Controller Update (KB2677596) - http://catalog.update.microsoft.com/v7/site/Search.aspx?q=2677596

Additional System Center 2012 and Private Cloud Resources

I hope you found this installment useful and please don't hesitate to let me know if elements within this conversation you'd like to discuss at greater depth.

On a loosely related note, over the past few months, I have written many articles on System Center 2012 and MS private cloud solutions. Links to a few of these resources are shown below.

Implementing rapid scale-out of a machine tier in VMM 2012 via PowerShell

Cloud: A quick note on accurately measuring Hyper-V host performance and utilization

What’s new in the Cloud Services Process Pack (CSPP) Release Candidate for System Center 2012?

Private Cloud in the real world: 5 lessons from the private cloud fabric

How to prepare for Orchestrator? Learn Opalis in a month of Friday afternoons! (free online resources)

PowerShell Scripts for System Center (Master Collection)

[Private Cloud]: Configuring anti-affinity between VMs on Hyper-V 2008 R2 SP1 and VMM 2012

SCVMM Tip: Testing PRO Integration in VMM / OpsMgr 2012

4 Hyper-V performance testing tips for the private cloud

Hyper-V Network Optimizations and Best Practices [Private Cloud]

Streamlining Linux VM configuration on Hyper-V with the Set-LinuxVM PowerShell module

Top 10 Features of Windows 8 (now 2012) [Infographic]

Tommy Gunn — Fri, 27 Apr 2012 08:03:14 GMT

A bit about what’s shiny and new in Windows 2012 via TinyHacker

My wish list for the Cloud Services Process Pack RTM

Pete Zerger — Thu, 26 Apr 2012 04:09:49 GMT

As someone who deploys Microsoft private cloud solutions in the real world, the RC release of the Cloud Services Process Pack solution accelerator was in my opinion a bit off target for what was really needed in a self-service portal solution for private cloud…particularly given the capabilities of the underlying platform, Service Manager 2012. But sometimes great things come from humble beginnings , so I have my fingers crossed.

In any event, there a few common pain points I hear when working with customers and private cloud adoption, many of which I think can be addressed in the RTM release the Cloud Services Process Pack (CSPP). But just to be safe, on the plane the other day I threw together my wish list for the RTM release of the Cloud Services Process Pack.

Dear ~~Santa~~ Microsoft,

Please find here my wish list for the RTM release of the Cloud Services Process Pack.

A Simple and Intuitive Self-Service Experience
The "Paradox of Choice" can turn users off of a solution very quickly. One of the complaints I hear about other self-service portals (e.g. - vCloud Director, IBM, etc.) is that they contain too many choices for users, resulting in confusion. An experience as simple as opening a helpdesk ticket, complete with familiar terminology, menu-driven choices is a must. A simple, plainly worded, intuitive self-service experience is paramount the success and perhaps the most important characteristic of a successful self-service solution (IMHO).

Closed Loop Change
At the end of the day, I want my private cloud solution closely aligned with my organizations ITIL processes. To fully get IT out of the way, the solution should include proper notification of the result of the request. With Service Manager 2012 at the core, properly skilled datacenter administrators should be able to implement this requirement with relative ease. It’s great automate the provisioning process and show the result in the self-service portal, but it’s even better close the loop with an e-mail. This one is easy, it's mostly a matter of just teaching people how to configure notification in Service Manager 2012 .

Automated Service Provisioning and Updates
Automated is the key word here. Automated provisioning instead of a service request that simply documents the request for manual implementation will deliver on a couple of the core promises of private cloud - agility and economy. With Orchestrator 2012 and the System Center integration packs as the core components for automated provisioning, some level of automation for service provisioning should be feasible. This is going to require either some free text entry from the user or fairly strict naming conventions within VMM service templates and drop-down menus (possibly setup by the IT org) to work within the capabilities of the Service Manager 2012 Self-Service Portal...but it is definitely possible.

Note: One of the primary challenges in automating self-service updates to a service created from template (while maintaining an intuitive user experience is due to the absence of machine tiers as CI’s in the CMDB. This is an irritation of mine right now as I work through development of custom solutions in VMM / SCOrch / OpsMgr / SvcMgr 2012 in my spare time. I do think automated updates ( machine tier scale out, minor VM changes, etc.) should be achievable with some work.

Automated Lifecycle Management

One of the most common issues within the enterprises VM sprawl, especially in development and test environments where nobody's keeping really close track. At the same time , test and development environments are where many enterprises are incorporating Hyper-V for the first time. When we automate provisioning in those environments and take IT out of the way, we create the potential for massive VM sprawl if we don't incorporate some measure of lifecycle management in the process. We saw elements of this in the RC release, and I think this needs to be carried over into proper reports that we can schedule and e-mail or as part of a dashboard. With great notification capabilities in Service Manager 2012, long as we see the expiration date in the CMDB we should be able sort of notification and de-provisioning workflow.

A Proper Chargeback \ Showback Dashboard
This is the item on my list I'm most worried about and a very important component of the right solution. The showback dashboard delivered in the RC fell short what was needed, and I'm sure the community let Microsoft know this . For anyone who tried the CSPP RC, I’d imagine this was a bit of a sore spot. With elements of cloud capacity built into the System Center 2012 framework, there should be no excuse for pushing this necessary feature off to the partner community. Adding cost center or other ownership monikers to VMs and service.

So please ~~Santa~~ Microsoft, deliver a proper chargeback dashboard for RTM

Additional Resources

Here are just a few additional private cloud-related posts featuring System Center 2012 and related technologies that may be of interest

Implementing rapid scale-out of a machine tier in VMM 2012 via PowerShell

Cloud: A quick note on accurately measuring Hyper-V host performance and utilization

What’s new in the Cloud Services Process Pack (CSPP) Release Candidate for System Center 2012?

Private Cloud in the real world: 5 lessons from the private cloud fabric

How to prepare for Orchestrator? Learn Opalis in a month of Friday afternoons! (free online resources)

PowerShell Scripts for System Center (Master Collection)

[Private Cloud]: Configuring anti-affinity between VMs on Hyper-V 2008 R2 SP1 and VMM 2012

SCVMM Tip: Testing PRO Integration in VMM / OpsMgr 2012

4 Hyper-V performance testing tips for the private cloud

Hyper-V Network Optimizations and Best Practices [Private Cloud]

Streamlining Linux VM configuration on Hyper-V with the Set-LinuxVM PowerShell module

#MMS2012 DVDS have been release!! well Almost.. #MMSFUN

Larry — Wed, 25 Apr 2012 16:22:38 GMT

I started looking into getting these files for my team, then I thought "too much work for for only these Turkeys to see..." Hopefully this will be bullet proof enough that it will just work..

Included is a script and Excel Spreadsheet that is used to grab all the MMS2012 Video presentations and copy them by category to the folder of your choice. Read the notes, as there is one line that needs to be modified so the files are downloaded to the correct folder.

PowerShell Snippet: Check for existence of a virtual machine

Pete Zerger — Wed, 25 Apr 2012 16:21:40 GMT

Just had a request for a code snippet to determine if a virtual machine exists via VMM Command Shell. Here is a quick IF-THEN-ELSE statement, which can be incorporated into a larger script or a function.

It basically consists of attempting to retrieve the VM and assigning to a variable, then determining if that variable is NULL or not. If it’s NULL, we know that no VM exists by that name.

$VM = "DC"

$VMName = Get-VM -name $VM

if (!$VMname) {

Write-Host "No VM named $VM exists" }

else

{ Write-Host "A VM named $VM already exists" }

Additional Resources

Many more samples on the Master List of System Center PowerShell

MMS 2012 Keynote 4/17: Summary for Geeks (#SYSCTR, #MMS2012)

Cameron Fuller — Tue, 17 Apr 2012 16:51:57 GMT

The first keynote of MMS 2012 represented the 10 year anniversary for MMS! Specific points from the keynote include the following items:

Announcements:

System Center 2012 was officially announced: A USB Key provided for MMS attendees with all of the System Center 2012 software

Windows Server 2012 name announced!
- (Previously referred to as “Windows Server 8”)
- This implies that Windows Server 2012 will be released in 2012
- Presenters discussed that System Center will be updated to support Windows Server 2012

Other items of note:

An SCVMM report showing included forecasting (shown below)

There was a great demo video from Vijay Tewari to deploy a bare metal host in SCVMM (Having met Vijay in the past I loved the coffee jokes integrated throughout- he may be the only guy at Microsoft I know who drinks more coffee than I do)
Demonstration of Operations Manager showing service level integration and dashboards in OpsMgr 2012, and .NET monitoring including application failures.
MCSE re-invented for the cloud: MCSE Private Cloud.
Windows Server 2012:
- No shared storage, migration of VM’s between servers. “Shared nothing live migration”
- Network virtualization in Windows Server 2012
Catapult’s own Jason Sharp was on the intro-video! (see the picture in the top left)

3 Tips to Streamline Application Configuration, Deployment and Troubleshooting in VMM 2012 service templates (PCITRW)

Pete Zerger — Wed, 11 Apr 2012 22:23:13 GMT

As part of the private cloud in the real world (PCITRW), I wanted to quickly share a some tips for building out and testing service templates in Virtual Machine Manager 2012 (VMM) in which you are deploying multiple applications. The application deployment functionality in VMM 2012 service templates is really great, but can be challenging when you are deploying a complex service with multiple applications. Below are three tips I can offer from my experience deploying services with VMM 2012 in the lab and in the field.

VMM 2012 includes three 1st class application deployment technologies, specifically:

WebDeploy – For IIS applications
Server App-V – To virtualize applications that can be virtualized (similar in many ways to App-V as you know it on the client side)
SQLDAC – SQL data tier applications for deployment and configuration of SQL databases.

And for everything else, we can use generic command executions (GCE), which generally come in the form of PowerShell, Windows Installer and batch files.

Here on some thoughts on how to ease the configuration and testing process.

1. Install Apps Manually First to Identify Role and Feature Dependencies

When you install an application for the first time, it’s a great idea to do it manually first just to determine if your VM template has all the prerequisite software components. If there are role and feature dependencies, these can be enabled in the properties of the VM template within the machine tier (see figure 1 below) , but there may be other components you need to install as well.

figure 1. Enabling automatic installation of roles and features in VMM service template to facilitate application deployment.

2. Configure Logging for Your GCEs

More often than not, a GCE is going to be a PowerShell script, Windows installer package (MSI) or batch file. Enabling logging is really important if you want to get to root cause in the event of a failure.

For example, verbose logging in an Windows installer package looks like this:

Msiexec /i "path\MSI\IBM WebSphere MQ.msi" /l*v c:\install.log /m mif_file
  TRANSFORMS="1033.mst" AGREETOLICENSE="yes" ADDLOCAL="Server"

And in the UI, if you click on the Advanced button (shown in figure 2), which opens the Advanced Script Command Settings window:

figure 2. Advanced Script Command Settings button

You can configure logging location and exit code match (for success or failure) in the UI (shown in figure 3):

figure 3. Advanced Script Command Settings detail

3. Build and Test Service Templates in Layers

If you are building a multi-tier service template, you may consider testing one tier at a time. For example, build the data tier and make sure your SQL install goes as planned. Then, make a copy of the template and add the application tier, and on a later iteration, the web tier. I like this approach particularly in situations where I am deploying several applications within the tiers of the service template.

Additional Resources

Here are a few additional resources related to Microsoft virtualization and private cloud you may find interesting.

MMS 2012-Birds of a Feather sessions: New to Microsoft System Center 2012 (#SYSCTR, #MMS2012, #MMSBOF12, #MMSBOF13)

Cameron Fuller — Tue, 10 Apr 2012 22:05:33 GMT

I just finalized on planning to moderate two sessions on “New to Microsoft System Center 2012” at MMS 2012. With all of the new features coming to the System Center 2012 stack these should be great discussions!

Abstract: “The long-awaited System Center 2012 product release is here! In this Birds-of-a-Feather discussion we will have an open forum focusing on new features available in System Center 2012 components including Configuration Manager/Endpoint Protection, Operations Manager, Service Manager, Data Protection Manager, Virtual Machine Manager, App Controller and Orchestrator. Additional topics include how the System Center 2012 stack works better together and how we can automate common processes using workflows in System Center.”

If you are attending MMS and interested in System Center please stop by and join this discussion (BOF12, BOF13)!

Thanks you to both Microsoft and one extremely cool colleague of mine out there (you know who you are!) for inviting me to moderate these sessions!

Tablet-a-day giveaway at MMS 2012 courtesy of System Center Central and the Inside MMS Podcast

Pete Zerger — Tue, 10 Apr 2012 00:15:55 GMT

I wanted to take a minute to let you know a bit about what we have planned for MMS 2012… whether you are attending or not.

Tablet-a-Day Giveaway

First, as a way of thanking our great community supporters and our rock star founding sponsors Quest and Veeam,, We’re going to be giving away a Kindle Fire every day of the conference!

For those of you who couldn’t make it, the Friday drawing is for you!. We’ll be sharing details of how you can enter to win the final Kindle Fire drawing without getting out of your chair!

Winners will be announced every day on the Inside MMS podcast, which will be broadcasting live from the Venetian bringing you a daily recap of big announcements, interviews with Microsoft Program Mangers and MVP subject-matter experts.

Big Tech Book Giveaway

Second, as a community with authors on multiple System Center and related technical books, we’ll be giving away more than 20 technical books, including the PowerShell 2.0 Bible, Opalis Unleashed and others.

Details

Stand by for contest details and more announcements as the conference approaches. Thanks again to our founding sponsors Quest and Veeam and to you, the System Center community. We look forward to talking to you in (or remotely from) MMS 2012 in Las Vegas!

Are you looking to participate in the System Center community via Twitter?

rorymccaw — Fri, 06 Apr 2012 03:27:43 GMT

If so, consider using the following hash tags to have your content followed by the greater System Center community at large.

Component	Hashtag
System Center all up or broad awareness (should be included in every tweet related to SC)	#sysctr
These are only included in tweets that are specific to a particular component:
Virtual Machine Manager	#scvmm
Service Manager	#scsm
Operations Manager	#scom
Data Protection Manager	#scdpm
Orchestrator	#scorc
Advisor	#scadvisor
App Controller	#scappctrl
Configuration Manager	#configmgr

Have fun, learn System Center. www.infrontconsulting.com/training.php

Error installing Server App-V PowerShell Cmdlets

Pete Zerger — Thu, 05 Apr 2012 18:13:53 GMT

I ran into a small roadblock yesterday when working with Server App-V I felt was worth sharing.

If you attempt to install Server App-V PowerShell cmdlets (for Server App-V agent or sequencer) on a clean Windows 2008 Server, you may get the following generic error screen.

Locating Error Source

If you look in the Application Windows Event Log, you can see the source of the error. In this case, a missing prerequisite of .Net Framework 3.5 not mentioned by the installer

Loading the Cmdlets

Once you install the cmdlets, whether on a computer with the Server App-V agent installed or with the Server App-V Sequencer installed, you can use import-module at a PowerShell prompt to verify availability.

On a Server App-V Agent computer

type import-module ServerAppVAgent

Then type get-command –module ServerAppVAgent | select name

On a Server App-V Sequencer

type import-module ServerAppVSequencer

Then type get-command –module ServerAppVSequencer | select name

Additional Resources

Here are a few additional resources related to Microsoft virtualization and private cloud you may find interesting.

System Center 2012 Technical Documentation Downloads

Pete Zerger — Wed, 04 Apr 2012 18:35:01 GMT

Here’s the full collection of System Center 2012 documentation available on TechNet, which as of this minute includes all but App Controller, Configuration Manager and Data Protection Manager.

System Center 2012 - Virtual Machine Manager: http://www.microsoft.com/download/en/details.aspx?id=6346

System Center 2012 - Service Manager: http://www.microsoft.com/download/en/details.aspx?id=27850

System Center 2012 - Operations Manager: http://www.microsoft.com/download/en/details.aspx?id=29256

System Center 2012 - Orchestrator: http://www.microsoft.com/download/en/details.aspx?id=29258

System Center 2012 - Unified Installer: http://www.microsoft.com/download/en/details.aspx?id=29257