ptg6432687 230 7 Optimizing the Hyper-V Host Server and Guest Sessions . Equal per IIS Application Pool—Allocates resources equally among all Internet Information Services (IIS) application pools, preventing one session from consuming all available CPU and memory resources. NOTE WSRM policies are only enforced when CPU usage climbs above 70%. The WSRM poli- cies are never active on processes owned by the core operating system or any items in the exclusion list. TIP Memory limits should be applied in policies only when the application, service, or process is having issues or not allocating memory usage properly on its own. A common task performed in WSRM is to create matching criteria rules. Matching criteria rules allow an administrator to define (or exclude) processes, services, or applications that should be monitored by WSRM. This definition is used later in the WSRM management process. To create a matching criteria rule, complete the following steps: 1. Launch Windows System Resource Manager by clicking Start, All Programs, Administrative Tools, Windows System Resource Manager. 2. Right-click the Process Matching Criteria item in the WSRM console and select New Process Matching Criteria. 3. Enter a unique name for the matching criteria in the Criteria Name box at the top and click Add under the Rules section. a. Enter the processes, services, or applications in the Included Files or Command Lines section of the Files or Command Lines tab. Or b. Select the object type (process, service application, or IIS application pool) from the drop-down list, and click the Select button and select the policy to apply. 4. To exclude items from the policy, check the Excluded Files or Command Lines check box. a. Enter the processes, services, or applications in the Included Files or Command Lines section of the Files or Command Lines tab. Or b. Select the object type (process, service application, or IIS application pool) from the drop-down list, and click the Select button and select the policy to apply. 5. Repeat the preceding steps to add all the exclusions and items that should be man- aged by or excluded from a WSRM policy. Download at www.wowebook.com ptg6432687 231 Using Capacity-Analysis Tools 7 Another task that is commonly performed is creating custom resource-allocation policies. Similar to “matching criteria rules” that look for specific process, service, and application criteria, the custom resource-allocation policy enables the administrator to define how much of a resource should be allocated to a specific process, service, or application. As an example, if only 20% of the system processing should be allocated to a print process, the resource allocation would be defined to limit the allocation of resources to that process. To create a custom resource allocation policy, complete the following steps: 1. Launch Windows System Resource Manager by clicking Start, All Programs, Administrative Tools, Windows System Resource Manager. 2. Right-click the Resource Allocation Policies option in the WSRM console, and select New Resource Allocation Policy. 3. Provide a name for the policy, and click the Add button in the Allocate These Resources section. 4. On the General tab, select the Process Matching Criteria and specify the percentage of processor time that will apply. 5. On the Memory tab, specify the maximum committed memory and working set limits. 6. The Advanced tab allows you to select which processors the policy should be assigned to and to suballocate processor resources. If you want to edit these parame- ters, make the changes and click OK. 7. Click OK when you have finished. The calendar component of WSRM can be used to schedule policy enforcement on a reglar basis and by one-time or recurring events. For example, policy enforcement might be necessary only during normal business hours. Calendar control is disabled by default and can be activated by right-clicking the Calendar item in the WSRM console and selecting the Enable or Disable option. To create calendar items based on scheduled times, complete the following steps: 1. Launch Windows System Resource Manager by clicking Start, All Programs, Administrative Tools, Windows System Resource Manager. 2. Expand the calendar item in the WSRM console by clicking the plus sign. 3. Right-click the Schedule option and select New Schedule. 4. Enter a name and description for the schedule. 5. Double-click a time slot in the New Schedule window, specify the policy, start time, and stop time, and then click OK. Instead of creating a calendar item based on scheduled times, you can create the calendar item based on a specific triggered event. To create calendar items based on specific events, complete the following steps: 1. Launch Windows System Resource Manager by clicking Start, All Programs, Administrative Tools, Windows System Resource Manager. 2. Expand the calendar item in the WSRM console by clicking the plus sign. 3. Right-click the Calendar Event option, and select New One Time Event. Download at www.wowebook.com ptg6432687 232 7 Optimizing the Hyper-V Host Server and Guest Sessions 4. Enter a name for the event. 5. Select Policy Name or Schedule Name, and select the appropriate policy. 6. Specify a start and end date and time (not available if associated with a schedule), and then click OK. For calendar events that you want to trigger based on recurring events, a rule can be created for this to happen. To create recurring events, complete the following steps: 1. Launch Windows System Resource Manager by clicking Start, All Programs, Administrative Tools, Windows System Resource Manager. 2. Expand the calendar item in the WSRM console by clicking the plus sign. 3. Right-click the Calendar Event option, and select New Recurring Event. 4. Enter a name for the event. 5. Select Policy Name or Schedule Name, and select the appropriate policy. 6. Specify a start and end time and specify a recurrence schedule, such as every Monday (not available if associated with a schedule), and then click OK. One example of where WSRM is useful is when an administrator wants to allocate system resources to sessions or users who are active on a Hyper-V host system. Configuring a WSRM policy for Hyper-V can ensure the sessions will not behave erratically and system availability will be stabilized for all the guest sessions hosted by the Hyper-V server. This is accomplished using the Equal per Session policy templates provided with WSRM or specif- ically allocating performance for each guest session, as covered in the “Accessing Hyper-V Resource Control” section of this chapter. To allocate resources to a Windows 2008 Terminal Services system, complete the following steps: 1. Launch Windows System Resource Manager by clicking Start, All Programs, Administrative Tools, Windows System Resource Manager. 2. Expand the Resource Allocation Policies option in the WSRM console, and select New Resource Allocation Policy. 3. Right-click Equal per Session or Equal per User, and select Set as Managing Policy. 4. A dialog box opens indicating that the calendar function will be disabled; click OK. 5. Click OK. Microsoft Baseline Configuration Analyzer (MBCA) The Microsoft Baseline Configuration Analyzer (MBCA) is a tool that uses a preestablished baseline to analyze a system. The MBCA is run from a command line and includes several switches to further customize its use. The gathered data is output into a report that can be reviewed to identify which baselines match and do not match. Baseline values are typi- cally based on best practices for system configuration, security, and so on. Systems will automatically download new baseline models as they are made available on the Microsoft Update site. You can download the MBCA from the System Tools section of the Microsoft download site at www.microsoft.com/downloads/. Download at www.wowebook.com ptg6432687 233 Using Capacity-Analysis Tools 7 Assessment and Planning Solution Tool The Assessment and Planning Solution tool provides a solution to IT personnel when faced with questions such as “Which product should we buy or deploy?” or “Are we ready for Windows 2008?.” Granted, there are multiple approaches to tackling questions like this; however, Microsoft has again developed a tool that will do most of the work for you. The Assessment and Planning Solution tool inventories and assesses systems, hardware, and software and makes product and technology recommendations based on those results. You can download the Assessment and Planning Solution tool from the Microsoft down- load site at www.microsoft.com/downloads/. System Center Capacity Planner (SCCP) 2007 System Center Capacity Planner 2007 is a tool for IT staff to plan their migration or deployment of System Center Operations Manager (OpsMgr) 2007 and Exchange 2007. SCCP can determine and recommend the necessary changes for deploying Exchange 2007 and OpsMgr 2007. This includes in-depth analysis of hardware, network architecture, placement of servers, and much more. SCCP 2007 can even advise on changes after deployment, whether they are planned or unplanned such as the addition of new users, new features such as Outlook Web Access (OWA), or changes to the network. You can find more information about SCCP 2007 at www.microsoft.com/systemcenter/sccp/default. mspx. System Center Operations Manager 2007 OpsMgr 2007 has replaced its popular predecessor, Microsoft Operations Manager (MOM). OpsMgr 2007 is a comprehensive monitoring and reporting solution that reports on conditions related to services, system, and network performance. In addition, it alerts administrators when problems arise, (for example, when critical services have failed to start, when CPU usage consistently stays above a designated threshold, or when excessive paging is observed by the OpsMgr agent). OpsMgr integrates directly with Active Directory, Windows 2008, and most other Microsoft technologies to provide an overall solution to help automate monitoring of critical systems and processes. OpsMgr uses management packs specific to the technology, such as Exchange 2007 or IIS 7.0, so little configuration is needed out of the box. Third-Party Toolset Without a doubt, many third-party utilities are excellent for capacity-analysis and perfor- mance-monitoring purposes. Most of them provide additional functionality not found in Windows 2008 Performance Monitor and other tools, but they have a cost and might have special requirements for deployment and integration into the organization’s network. You might want to evaluate some third-party utilities to get a more thorough understanding of how they might offer more features than Microsoft solutions. Generally speaking, these utilities enhance the functionality inherent to Microsoft monitoring solu- tions, such as scheduling, an enhanced level of reporting functionality, superior storage capabilities, the ability to monitor non-Windows systems, or algorithms for future trend analysis. Table 7.2 lists some of these third-party tools. Download at www.wowebook.com ptg6432687 234 7 Optimizing the Hyper-V Host Server and Guest Sessions Although it might be true that most third-party capacity-analysis and performance-moni- toring products might add more or different functionality to your capacity-analysis and performance-monitoring procedures or goals, there are still pros and cons to using them over the free tools included with Windows 2008 or other solutions available from Microsoft. The key is to decide what you need to adequately and efficiently perform capacity-analysis and performance-monitoring procedures in your environment. Taking the time to research and experiment with the different solutions available today, from Microsoft and others, will only benefit you in making an informed decision for managing your Windows 2008 environment. Optimizing the Performance of Hyper-V Host Servers and Guest Sessions With the various tools covered in the section “Using Capacity-Analysis Tools,” baseline system performance and capacity analysis will have been identified. The next step is to get into the Hyper-V host and guest sessions and tune the guest sessions so that performance can be directly managed and optimized. The two key areas of resource management that come in focus are guest session resource allocation and guest session disk image optimization. Resource Allocation to Hyper-V Guest Sessions For each Hyper-V guest session, a setting allows the allocation of Hyper-V host resources to the guest session. The more resources allocated to a guest session, the more resource demands the session can make on the resource pool (and thus the more scalable impact one session could have on another session). Accessing Hyper-V Resource Control Resource Control in Hyper-V is allocated in the Virtual Processor section of the settings for each guest session. To go to the Virtual Processor section of a guest session, complete the following steps: TABLE 7.2 Third-Party Capacity-Planning and Monitoring Tools Utility Name Company Website AppManager Suite NetIQ Corporation www.netiq.com/products/per formancemgmt/ BMC Performance Manager BMC Software www.bmc.com/ HP OpenView HP www.openview.hp.com/ Robomon Heroix www.robomon.com/ UniCenter NetCenter CA www.ca.com/ Download at www.wowebook.com ptg6432687 235 Optimizing the Performance of Hyper-V Host Servers and Guest Sessions 7 FIGURE 7.15 Resource Control settings. 1. Launch the Hyper-V Manager tool. 2. Right-click a virtual machine that you want to view or edit the Resource Control of and choose Settings. 3. In the leftmost pane, click Processor. Notice the Virtual Processor screen on the right side, and note the Resource Control section, as shown in Figure 7.15. Identifying Hyper-V Resource Control Settings In the Virtual Processor settings, three settings are available to configure resource alloca- tion for the guest session: . Virtual Machine Reserve (Percentage)—This number (default = 0, but can range from 0 to 100) is how much of the host server’s available resources the guest session should be guaranteed. If the machine reserve is set to 100, this guest session will get 100% of the resources of the server, and thus no other guest session will be able to run on this server. Consider this number as a “minimum” amount of resources the guest session should receive, whether it needs that much or not. . Virtual Machine Limit (Percentage)—This number (default = 100, but can vary from 0 to 100) is the percentage maximum that this guest image will take of system resources. When set to 100, it is possible for this guest image to take up 100% of the resources of the system. If you want a guest image to take up no more than 25% of the resources of the host system at maximum, you should change this number to 25. Remember, however, if you set the limit at 25 and the host server is underutilized Download at www.wowebook.com ptg6432687 236 7 Optimizing the Hyper-V Host Server and Guest Sessions and this guest session can use extra resources, the guest session will max out at 25% despite the guest session’s need or the host server’s capacity to provide the resources required. . Relative Weight—The relative weight (default = 100) is used when a contention exists between guest sessions on resource control. If three guest sessions are all trying to utilize 100% of the host server resources, based on the relative weight (where the higher number is given preference over a guest session with a lower number), the server with the highest number will take a higher priority in resource allocation than a guest session with a lower relative weight number. Setting Hyper-V Resource Control Options After deciding how you want to allocate resources to each guest session on the host system, you configure the settings one by one for each guest session. This includes keying in the requested value for each of the three settings, and then clicking OK to set the sessions for the guest session. Repeat this process for all guest sessions running on the host server system. Optimizing Disk Configuration for Hyper-V Guest Sessions Hyper-V guest session disk configurations can significantly impact performance of the guest session and how that guest session impacts the overall performance of the host server. Two factors should be considered relative to disk configuration: the type of disk selected, and where the disk image is stored. Choosing Disk Type for Performance Considerations There are two major differences in disk types in Hyper-V, one is a dynamically expanding disk, and the other is a fixed-size disk. (I note a third in the following bulleted list.) This disk type is selected at the time the guest image is created; by default, the image type selected is a dynamically expanding image type. After selecting a disk type and installing the OS on the image, you cannot easily switch to a different disk type. However, it is not impossible to convert a dynamically expanding to fixed or vice versa. At the time of installation, when you are creating a new virtual hard disk (VHD), you are given the option (as shown in Figure 7.16) to choose a dynamically expanding disk type, fixed-size disk type, or a differencing disk type. The differences are as follows: . Dynamically Expanding—A dynamically expanding disk type allows a disk image to start off as small as possible, and the image grows (up to a maximum size defined at the time of configuration; default = 127GB). The advantage of a dynamically expanding disk type is that it takes up little room on the disk (because the image may take up only 2GB or 4GB of space to start). As the image demand grows, the size of the image file grows. . Fixed Size—The fixed-size disk is one where the size of the disk image is selected at the time of installation, whether that is 10GB or 20GB or 100GB. The disk space is immediately allocated regardless of whether the guest session uses that amount of space. So, if 100GB is allocated to the fixed size of a guest image, the guest image will take up 100GB of disk space. Download at www.wowebook.com ptg6432687 237 Optimizing the Performance of Hyper-V Host Servers and Guest Sessions 7 FIGURE 7.16 Choosing the disk type. . Differencing—The differencing disk type is one where an existing image already exists, and only changes specific to this guest session image are stored for the guest session. This option is used for servers that are pretty much identical and not likely change much. The organization can reduce the amount of disk space taken up by having multiple servers each taking up 2GB or 4GB of space when a single image of, say, 2GB can be shared by multiple guest sessions. NOTE This VHD Configuration Wizard comes up only when creating a new VHD, not when cre- ating a new guest image. When creating a new guest image, the default disk image type is set to dynamic if you choose New Disk. If you choose an existing virtual disk, you can choose a fixed disk that has already been created via the VHD Configuration Wizard. As stated previously, a dynamically expanding disk type allows for a guest image to start off small and grow as needed. From a performance perspective, however, every time the image needs to grow, it has a drastic impact on the guest session and the host server. Fixed-size disks images take up a static amount of disk space whether the space is needed or not. And even though the space might not be needed, at least the image will not grow during the production day and cause performance impact on the server. So, for the best performance, choose the fixed-size disk type. Download at www.wowebook.com ptg6432687 238 7 Optimizing the Hyper-V Host Server and Guest Sessions Choosing the Location for Disk Images for Performance Considerations The other key factor when considering performance on a Hyper-V host system is the storage location of disk images. If disk images are stored all on the C: drive of the host server, and the disk is not a fast disk, all the guest images shairng a single slow disk are impacted. If the guest images are stored on a remote disk subsystem, such as an iSCSI storage area newtork (SAN) server, the communication speed between the Hyper-V host and the iSCSI SAN is not that fast, and so again, there is a performance impact on guest session operation. If the guest sessions are stored on a high-speed FibreChannel SAN with extremely high read/write performance (both on the disk and the connectivity between the host and the high-speed SAN), however, the performance of the guest images is enhanced. Take great care when considering where to store disk images. Your decision will impact the overall performance of the guest images and the host server managing the guest image sessions. Monitoring System Performance Capacity analysis is not about how much information you can collect; it is about collect- ing the appropriate system health indicators and the right amount of information. Without a doubt, you can capture and monitor an overwhelming amount of information from performance counters. There are more than 1,000 counters, so you’ll want to care- fully choose what to monitor. Otherwise, you might collect so much information that the data will be hard to manage and difficult to decipher. Keep in mind that more is not necessarily better with regard to capacity analysis. This process is more about efficiency. Therefore, you need to tailor your capacity-analysis monitoring as accurately as possible to how the server is configured. Every Windows 2008 server has a common set of resources that can affect performance, reliability, stability, and availability. For this reason, it’s important that you monitor this common set of resources, namely CPU, memory, disk, and network utilization. In addition to the common set of resources, the functions that the Windows 2008 server performs can influence what you should consider monitoring. So, for example, you would monitor certain aspects of system performance on file servers differently than you would for a domain controller (DC). Windows 2008 can perform many functional roles (such as file and print sharing, application sharing, database functions, web server duties, domain controller roles, and more), and it is important to understand all those roles that pertain to each server system. By identifying these functions and monitoring them along with the common set of resources, you gain much greater control and understanding of the system. The following sections go more in depth on what specific items you should monitor for the different components that constitute the common set of resources. It’s important to realize, however, that several other items should be considered regarding monitoring in Download at www.wowebook.com ptg6432687 239 Monitoring System Performance 7 addition to the ones described in this chapter. Consider the following material as just a baseline of the minimum number of things to begin your capacity-analysis and perfor- mance-optimization procedures. Key Elements to Monitor for Bottlenecks As mentioned, four resources compose the common set of resources: memory and pagefile usage, processor, disk subsystem, and network subsystem. They are also the most common contributors to performance bottlenecks. A bottleneck can be defined in two ways. The most common perception of a bottleneck is that it is the slowest part of your system. It can either be hardware or software, but generally speaking, hardware is usually faster than software. When a resource is overburdened or just not equipped to handle higher work- load capacities, the system might experience a slowdown in performance. For any system, the slowest component of the system is, by definition, considered the bottleneck. For example, a web server might be equipped with ample RAM, disk space, and a high-speed network interface card (NIC), but if the disk subsystem has older drives that are relatively slow, the web server might not be able to effectively handle requests. The bottleneck (that is, the antiquated disk subsystem) can drag the other resources down. A less common, but equally important form of bottleneck, is one where a system has significantly more RAM, processors, or other system resources than the application requires. In these cases, the system creates extremely large pagefiles and has to manage very large sets of disk or memory sets, yet never uses the resources. When an application needs to access memory, processors, or disks, the system might be busy managing the idle resource, thus creating an unnecessary bottleneck caused by having too many resources allocated to a system. Thus, performance optimization not only means having too few resources, but also means not having too many resources allocated to a system. Monitoring System Memory and Pagefile Usage Available system memory is usually the most common source of performance problems on a system. The reason is simply that incorrect amounts of memory are usually installed on a Windows 2008 system. Windows 2008 tends to consume a lot of memory. Fortunately, the easiest and most economical way to resolve the performance issue is to configure the system with additional memory. This can significantly boost performance and upgrade reliability. Many significant counters in the memory object can help you determine system memory requirements. Most network environments shouldn’t need to consistently monitor every single counter to get accurate representations of performance. For long-term monitoring, two very important counters can give you a fairly accurate picture of memory pressure: Page Faults/sec and Pages/sec memory. These two memory counters alone can indicate whether the system is properly configured with the proper amount of memory. Table 7.3 Download at www.wowebook.com . third-party tools. Download at www.wowebook.com ptg64 3268 7 234 7 Optimizing the Hyper-V Host Server and Guest Sessions Although it might be true that most third-party capacity-analysis and performance-moni- toring. needed out of the box. Third-Party Toolset Without a doubt, many third-party utilities are excellent for capacity-analysis and perfor- mance-monitoring purposes. Most of them provide additional functionality. ptg64 3268 7 230 7 Optimizing the Hyper-V Host Server and Guest Sessions . Equal per IIS Application Pool—Allocates resources equally among all Internet Information Services (IIS) application