Home > Articles > Operating Systems, Server > Microsoft Servers

Looking Inside Microsoft System Center Operations Manager 2007

  • Print
  • + Share This
This chapter explains the relationships between the computers functioning as components of a management group so that the job of deploying and supporting OpsMgr 2007 becomes easier and more effective.
This chapter is from the book

In This Chapter

  • Architectural Overview
  • Windows Services
  • Communications
  • How Does OpsMgr Do It?
  • Presentation Layer

Microsoft System Center Operations Manager 2007 (OpsMgr) is a monitoring and operations management system, implemented using one or more computers that perform their assigned roles as components of a management group. The components cooperate over several secure communication channels to achieve management information workflow and present information to operators and administrators. The most important data collected is the health of the managed objects; this health status is arrived at via models that affect the tactical placement of software probes called monitors.

This chapter endeavors to make these terms and relationships clear so that the job of deploying and supporting OpsMgr 2007 becomes easier and more effective. Those readers tempted to skip this chapter covering OpsMgr internals, definitions, and concepts are probably asking themselves, "What practical use can I expect to get from reading this chapter?" Some administrators avoid looking under the hood deliberately, and that's totally OK. For those individuals, we do recommend reading at least the "Management Group Defined" section of this chapter.

So, for those OpsMgr administrators who yearn to know exactly what is going on behind the scenes, this chapter is for you. We want you to understand the lingo and reasoning used by the software developers of Operations Manager. In doing so, we hope that more advanced material about OpsMgr will make sense more quickly to you, the OpsMgr administrator, when reading this book, using the product, or interacting with fellow professionals in the Microsoft systems management community.

Architectural Overview

This chapter looks at OpsMgr design and internals at two levels:

  • The macro level—We'll look at the computer roles that comprise a management group.
  • The micro level—We'll examine the objects that constitute a management pack, in particular its workflow and presentation of data to the operator.

As an OpsMgr administrator, you have no influence over server component characteristics—these are hard-coded features of the Operations Manager software and hardware architecture. On the other hand, administrators can enjoy almost complete flexibility regarding the manner in which management packs are utilized.

OpsMgr administrators of the smallest environments—administrators who will run all applicable OpsMgr components on a single server and manage only computers and devices on their local area network (LAN)—generally are less concerned about this section on OpsMgr architecture. In that small-scale scenario of the "all-in-one" management group server, there is much less to be concerned about with architectural considerations of the various OpsMgr computer roles (components) as long as you stay below capacity thresholds for that single server and its network segment. In this simplest OpsMgr environment, the only OpsMgr components not resident on the single server are the OpsMgr agents running on the managed computers on the network.

However, many OpsMgr administrators will need to distribute multiple components across different servers, deploying OpsMgr roles across multiple computers. Even OpsMgr 2007 deployments on small business networks may include an Audit Collection Services (ACS) Component to centralize security event auditing, or an OpsMgr Gateway Server Component to monitor service delivery at a branch office where there is no Virtual Private Network (VPN) connectivity. Deploying the feature sets added when installing these additional roles, by definition, adds one or more physical management servers to the management group and requires an understanding of Operations Manager 2007 management group architecture. Chapter 4, "Planning Your Operations Manager Deployment," and Chapter 5, "Planning Complex Configurations," provide information on hardware specifications and sizing server configurations.

Management Group Defined

A management group is an instance of the Microsoft end-to-end service management solution named Operations Manager 2007. Organizations may host several management groups (instances of OpsMgr on their networks) if appropriate for their business needs. Likewise, any managed computer or device can participate in one or more instances (management groups) of OpsMgr if appropriate. Most organizations of all sizes deploy a single management group, which is analogous to a single Active Directory (AD) forest or a single Exchange organization. Most organizations, including some very large ones, have their business needs met with just one AD forest and one Exchange organization.

Figure 3.1 illustrates a default, single management group in an organization, and contrasts that with a more complex implementation one might encounter in a large organization. In the simple all-in-one example on the left in Figure 3.1, all OpsMgr components are installed on one server, which is the only OpsMgr server in the single management group serving the managed computers (agents) in the organization. Several hundred computers can be managed with an all-in-one deployment of OpsMgr 2007.

Figure 3.1

Figure 1 Contrasting the smallest with a very large OpsMgr 2007 deployment model.

In the complex large organization scenario on the right-hand portion of Figure 3.1, a single computer agent is reporting simultaneously to two management groups (known as a multihomed agent), while one of those management groups, through its Root Management Server (RMS), participates with several connected management groups. This creates an architecture capable of servicing tens of thousands of widely distributed computers.

You will seldom need multiple management groups to get the most out of OpsMgr 2007 since the product's design provides full functionality to all but the largest of organizations while still using a single management group. For the very large organization (over 10,000 computers or over 100 remote sites), deploying several OpsMgr 2007 management groups can distribute the workload. Connecting these management groups enables you to query multiple management groups from the same Operations console.

Both having more than one production instance of OpsMgr in your organization and having a computer or device report status to more than one management group are advanced configurations to accomplish particular business goals. We describe these situations in Chapter 10, "Complex Configurations."

Server Components

Here are a dozen possible computer components, or roles, that can be deployed in an OpsMgr 2007 management group. Focusing now on what components constitute a single OpsMgr 2007 management group, let's begin with describing the core, or basic, server components. The core components are those that an OpsMgr 2007 deployment must include to have minimum functionality. These basic components (displayed in Figure 3.2) are installed in every management group, including the all-in-one server OpsMgr environment.

Figure 3.2

Figure 3.2 OpsMgr 2007 core components combined in one server, and distributed across dedicated servers.

  • Operations Database Server Component—The heart of the management group is the Operations database. The Operations database contains operational data about managed objects, the configuration store of what objects are managed, and all customizations to the OpsMgr environment. The Operations database is the central repository and processing point for all data in a management group. When you install an OpsMgr 2007 management group on multiple computer systems, the first thing to take place is installing the Operations database on an existing SQL 2005 database server running Service Pack (SP) 1 or later. The Operations Database Server Component can be clustered in high-availability environments.
  • Root Management Server Component—The first management server installed in a management group is the Root Management Server Component. Like all OpsMgr 2007 management servers, the RMS sends configuration information to managed computers and receives data from agents. The RMS alone runs some distinctive services that the entire management group depends on, and like the Operations Database Server Component, the RMS Component can be clustered. The RMS requires that the Operations database be available and accessible. The function of the entire management group also depends on the RMS; in high-availability environments you should consider clustering both the Operations database and the RMS components.
  • Agent Component—The Agent Component is used to monitor servers and clients. This is a Windows service that runs on managed servers and client computers. You might create an all-in-one server management group whose only purpose is monitoring network devices such as routers or switches; in that case, no agents need to be deployed. However, for most OpsMgr setups, the deployment of core management group components is not complete until one or more computers are selected for management and the Agent Component is installed. As we mentioned previously in the "Management Group Defined" section of this chapter, an OpsMgr 2007 agent can participate in more than one management group simultaneously.
  • Console Component—The OpsMgr 2007 console is the only application needed to interact with the management group, and it is used by both operators and administrators. Operations Manager 2007 implements role-based security to ensure an optimized experience for all users. There is also a web-based console with a subset of the regular console functions.

    Each console connects directly to the RMS, even if additional management servers have been deployed in the management group. This dependence makes RMS availability critical to perform almost every function in OpsMgr 2007. The first time a user opens the Operations console, there is a prompt to enter the name of the RMS, unless the user accessing the console is at a management server. After connecting, the console stores that server name, as well as the management group name, in the Connect to Server dialog box shown in Figure 3.3.

    Figure 3.3

    Figure 3.3 The Connect to Server dialog box stores the name of the RMS and associated management group.

The four components listed here are mandatory components and required for any OpsMgr management group to function. In addition, there are two core components related to reporting that most OpsMgr administrators will install regardless of their environment size:

  • Reporting Data Warehouse Server Component—A long-term data store is created with the Reporting Data Warehouse Server Component. The data warehouse stores aggregated historical performance and availability data beyond the few hours or days of data available in the Operations database. Without a data warehouse, an OpsMgr management group will only present information based on the real-time and very recent data captured in the Operations database, which is aggressively groomed of historical data. The Reporting Data Warehouse Server Component can be hosted on a clustered SQL Server backend.
  • Reporting Server Component—This component adds the reporting function to an OpsMgr management group and is required for the Reporting Data Warehouse Server Component. The Reporting Server is installed on a server running SQL Reporting Services 2005 SP 1 or later. Because of the integration between the Operations console and the Reporting Server, it is transparent to the user that the data for the reports is coming from the Reporting Server and not the Operations database or the RMS. This differs from the Microsoft Operations Manager (MOM) 2005 Reporting implementation.

    You can install the Reporting Data Warehouse Server and Reporting Server Components on the same Windows server, although in large and high-availability environments, these two components typically run on dedicated servers.

Finally, there are six optional components in an OpsMgr management group. Computers are deployed with these components as needed or desired to increase the monitoring capacity, or to add further features to the management group:

  • Management Server Component—This component refers to additional management servers installed after the RMS is installed. The primary reasons to deploy additional OpsMgr 2007 management servers are to enable agent failover and to manage a larger number of objects. There are specific procedures to promote a management server to the RMS Role in a disaster-recovery scenario, which we discuss in Chapter 12, "Backup and Recovery." You would also install an additional management server to host the Audit Collector Component, described later in this list, because that component requires installation on an existing OpsMgr management server that is not the RMS.
  • Audit Database Server Component—SQL Server 2005 is required for the Audit Database Server Component when adding the Audit Collection Services feature to the management group. Security events from managed computers are stored in this database and are used in generating reports. The Audit Database Server can be a clustered service for high availability. Reports on security events are generated from the Audit database.
  • Audit Collector Component—This server function collects events from the audit collection–enabled agents. The Audit Collector Component is added to an existing OpsMgr management server. Audit collection is enabled on OpsMgr agents by running a task in the OpsMgr console. Each collector needs its own individual Audit Database Server. The Audit database can be located on the same computer as the ACS Collector, but for optimal performance, each of these components should be installed on a dedicated server.
  • Web Console Server Component—Any OpsMgr management server running the Internet Information Services (IIS) web server service can optionally host a web-based version of the OpsMgr console. Functionally similar to using a thin client much like Outlook Web Access (OWA), operators can view topology diagrams and performance charts and run tasks made available to them appropriate for their role. The Web Console Server might be a management server dedicated to hosting this role in an organization that makes heavy use of the Web console.
  • Gateway Server Component—A communications conduit to monitoring agents in untrusted domains (or on remote networks without routed network connectivity), this server resides in an external environment and uses certificates to secure communication back to the other roles in the management group. A gateway server can also host the Audit Collector Component.
  • Client Monitoring Server Component—The Client Monitoring Configuration Wizard is used to configure the Client Monitoring Server Component on one or more management servers in a management group. The Agentless Exception Monitoring (AEM) Client Component is activated by a Group Policy Object (GPO) applied to client computers. An important note is that the management server and AEM clients must be in the same domain or fully trusted domains.

Figure 3.4 illustrates a management group with all components on distributed servers, and with many high-availability features deployed. This large-enterprise management group could provide end-to-end service monitoring of many thousands of objects with a high degree of reliability.

Figure 3.4

Figure 3.4 All basic and optional OpsMgr server roles deployed on dedicated servers.

  • + Share This
  • 🔖 Save To Your Account