3. Importing Management Packs
After the initial installation, OpsMgr includes only a
few management packs. The management packs contain all the discoveries,
monitors, rules, knowledge, reports, and views that OpsMgr needs to
effectively monitor servers and applications. One of the first tasks
after installing OpsMgr 2007 is to import management packs into the
system.
There are several management packs in the Internet
catalog on the Microsoft website. These include updated management
packs, management packs for new products, and third-party management
packs. Only load the management packs that are going to be used because
each additional management pack increases the database size, adds
discoveries that affect the performance of agents, and in general
clutters up the interface.
The key management packs for a Lync Server 2010 environment follow:
For each of these management packs, it is important
to load the relevant versions only. For example, if the environment
includes Windows Server 2008, load only the Windows Server Core OS 2008
management pack. If the environment includes both Windows Server 2003
and Windows Server 2008, load both the Windows Server Core OS 2003 and
the Windows Server Core OS 2008.
In versions of OpsMgr prior to R2, the management
packs had to be downloaded from the Microsoft website one by one, the
MSI installed one by one, and the management packs imported one by one.
Dependencies were not checked unless additional steps were taken to
consolidate the management pack files prior to importing. This was a
labor-intensive process. Also, there was no easy way to check for
updates to already installed management packs.
In OpsMgr 2007 R2, a new Management Pack Import
Wizard was introduced. This wizard connects directly to the Microsoft
management pack catalog and downloads, checks, and imports management
packs. It even performs checks to ensure that the management packs are
the latest versions. This is a huge improvement over the old method to
import management packs.
To import the key management packs, perform the following steps:
1. | Launch the Operations Console.
|
2. | Select the Administration section.
|
3. | Select the Management Packs folder.
|
4. | Right-click the Management Packs folder and select Import Management Packs.
|
5. | Click the Add button and select Add from Catalog.
|
6. | Click the Search button to search the entire catalog.
Note
The View pull-down in the Management Pack Import
Wizard includes four options, which are All Management Packs in the
Catalog, Updates Available for Installed Management Packs, All
Management Packs Released in the Last 3 Months, and All Management Packs
Released in the Last 6 Months. The Updates option checks against the
already-installed management packs and enables the download of updated
versions of these.
|
7. | Select the key management packs from the preceding list and click Add
for each of them. Each of the major management packs might include a
number of submanagement packs for discovery, monitoring, and other
breakdowns of functionality.
|
8. | After adding management packs, click OK.
|
9. | The
wizard now validates the added management packs, checking for versions,
dependencies, and security risks. It enables problem management packs
to be removed and dependencies to be added to the list.
|
10. | Click Install to begin the download and import process. Progress shows for each of the management packs imported.
|
11. | After all the management packs complete, click Close to exit the wizard.
|
After
the import completes, the management packs take effect immediately.
Agents discover based on the schedule specified in the management packs,
and monitors and rules deploy.
4. Deploying OpsMgr Agents
OpsMgr agents are deployed to all managed servers
through the OpsMgr Discovery Wizard, or by using software distribution
mechanisms such as Active Directory GPOs or System Center Configuration
Manager 2007. Installation through the Operations Console uses the fully
qualified domain name (FQDN) of the computer.
When searching for systems through the Operations
Console, you can use wildcards to locate a broad range of computers for
agent installation. Certain situations, such as monitoring across
firewalls, can require the manual installation of these components.
The Discovery Wizard can discover and configure
monitoring for Windows computers, UNIX/Linux computers, and network
devices. It pushes agents to Windows and UNIX/Linux computers, if the
proper rights are provided, such as an account with local administrator
rights or a root account.
To install domain member agents using the Discovery Wizard, perform the following steps:
1. | Launch the Operations Console and select the Administration section.
|
2. | Right-click the top-level Administration folder and select Discovery Wizard.
|
3. | Select the Windows computers and click Next.
|
4. | Select Automatic computer discovery and click Next. This scans the entire Active Directory domain for computers.
|
5. | Leave the Use selected Management Server Action Account and click Discover. This starts the discovery process.
|
6. | After
the discovery runs (this might take a few minutes), the list of
discovered computers displays. Select the devices that should have
agents deployed to them.
Note
The list includes only systems that do not already
have agents installed. If a computer has an agent installed, the wizard
excludes it from the list of devices.
|
7. | Click Next.
|
8. | Leave the Agent installation directory and the Agent Action Account at the defaults, and click Finish.
|
9. | The
Agent Management Task Status window appears, listing all the computers
selected and the progress of each installation. As shown in Figure 1, the LS1.companyxyz.com agent installation task starts.
|
10. | Click Close when the installation completes.
|
Even if the window is closed before the installs
complete, the results of the installs can be viewed in Task Status view
in the Monitoring section of the Operations Console.
The agent deployment is efficient, and a large number
of computers can be selected for deployment without any issues. The
agents start automatically and monitor as they are discovered.
After installation, it might be necessary to wait a
few minutes before the information from the agents is sent to the
management server.
During the next few minutes after installation, the
agent contacts the management server and establishes a mutually
authenticated, encrypted communication channel with the assigned
management server. If the agent was pushed through a software delivery
system such as System Center Configuration Manager 2007, the agent
determines the management server through Active Directory–integrated
discovery.
The agent downloads rules to discover the various
applications and components it’s hosting, enabling the correct
application-specific management packs to be applied. This discovery
process runs periodically to ensure that the correct rules are always
applied to the server.
5. SCOM Dip Stick Health Checks
When driving, the conscientious driver goes through a set of basic automobile checks, including the following:
These are sometimes referred to as the dip stick health checks because the automobile oil level is checked with a dip stick.
Like any other complicated technology, Operations
Manager 2007 can have problems in a variety of ways ranging from running
out of disk space, to failing to send email notifications, to having
agents stopped, and so forth. To make sure that Operations Manager is
functioning properly, a set of dip stick health checks can be performed
to make sure everything is running smoothly.
These are the tasks that the OpsMgr administrator
should do every day to verify the health and proper operation of the
OpsMgr infrastructure:
1. | Verify
that you have received notifications by e-mail. Confirm that you are
getting notifications within the normal range. Too many notifications is
a bad sign and too few (or none) is also a bad sign.
|
2. | Review
OpsMgr daily reports sent by e-mail or in the console. If using the
console, the reports are stored in the Favorites folder in the Reporting
space. |
3. | In the Operations Manager console, review the Active Alerts view. This shows you new alerts.
|
4. | In the Operations Manager console, review the All Alerts view. This shows you both new and closed alerts.
|
5. | In the Operations Manager console, review the Agent Health State view in the \Operations Manager\Agent node. Investigate Critical, Warning, or Not Monitored states.
|
6. | In the Operations Manager console, review the Active Alerts view in the \Operations Manager\Agent node. Investigate Critical or Warning alerts.
|
7. | In the Operations Manager console, review the Management Server Health State view in the \Operations Manager\Management Server node. Investigate Critical, Warning, or Not Monitored states.
|
8. | In the Operations Manager console, review the Active Alerts view in the \Operations Manager\Management Server node. Investigate Critical or Warning alerts.
|
After
reviewing these health check points, an administrator can be confident
that the Operations Manager 2007 R2 infrastructure is functioning
properly.
The second check recommends reviewing the daily
reports. The recommended Operations Manager health reports to review on a
daily basis are as follows:
Alert Logging Latency report— This report tells you the length of time between a raised event to a generated alert. This should be under 30 seconds.
Send Queue % Used Top 10 report—
This report tells you whether agents are having trouble uploading their
data to the management servers. These queues should be less than 1
percent.
Top 10 Most Common Alerts report— This report analyzes the common alerts that are generated and are good for identifying alert-tuning opportunities.
Daily Alert report—
This report gives you a complete list of alerts that were generated.
This is detailed, but is good for chasing down problems uncovered in
other checks.
These checks should give a good sense of the operational health of the SCOM infrastructure.
