3

I am trying to understand SCOM(OpsMgr) and I am little confused as to when the data is collected and the order of operations in the initial setup. At this point I have completed SCOM installation and added couple of agents. Here are my questions.

  1. What should be the order of setup for Agent Installation and adding Management Packs. Do I Install Agent first then add Management packs or Management packs first then agent.

  2. Is data collected as soon as the relevant Management Pack are added? Lets say I add an agent that has SQL Server or Exchange. If I add SQL Server Mgmt Pack, would SCOM immediately start collecting data from every server that has exchange/SQL Server installed?

    Is there a way I could specify not to collect SQL information from certain SQL Servers?

  3. How are the Groups defined in real world?. We have Linux,Windows Servers, SQL Servers, Oracle Servers, IIS, Apache, Network Adapters, Exchange, DNS, etc. It will helpful to know how this is setup in your company and the general idea to manage groups so that they are just enough not too many nor too little.

Community
  • 1
Afroz
  • 145
  • 6

1 Answers1

2

  1. It doesn't matter. MPs are pushed to agents very soon after they are imported. From my experience, most people install SCOM without any additional MPs at first, then add them in an orderly fashion. If you just install a whole slew of MPs at one time, you might get alert floods that are difficult to pinpoint. When installing MPs, be sure to read the documentation. Many MPs have additional setup steps in order to get them running properly. E.g. runas accounts often must be configured and distributed in order to collect data properly. It's also good to review all of the monitors and rules to be sure you've enabled what you need and disabled those you don't. The more actionable your alerts are, the happier you will be

  2. As soon as SCOM distributes the MPs to the agents, it begins collecting data. Each monitor/rule has its own configured frequency. Sometimes these are configurable; sometimes they are not. You'll have to look into each one to see.

  3. MPs often come with built in groups. For example, the SQL MP creates a "SQL Servers" group that will display every server with SQL installed. In my environment, I've taken that group and created my own based on it, then excluded servers that our DBAs aren't concerned with. Then I only send them alerts pertaining to servers in that group. I don't know much about other companies, but we use groups as organizational tools to control access to the SCOM console and to group overrides for specific servers or monitoring objects.

I hope that helps.

Czadd
  • 71
  • 3