Occasionally an Operations Manager 2007 management group might need to be completely recovered for some reason. This would be required in the case of unrecoverable systems due to database corruption, or complete loss of database and/or other key components, and there are no backups to recover these systems or components. I have found this happens more often in lab environments, where there is a lot of unmanaged testing and corruption is more likely to occur.
This article provides a high-level, step-by-step process to completely remove the core Operations Manager server roles and components in order to start fresh with a new installation of the management group. It is a two-phased approach, which includes decommissioning the current installation and then installing a new Management Group. The process outlined here offers a relatively short recovery time, since we will not need to re-install any agents on the computers that were previously managed.
I have been successful using this process on more than a couple occasions. However, these procedures have not been tested and approved by the product group. Using this guidance should be a last resort, after calling into Premier support and attempting to resolve any corruption issues (especially in a production environment).
Also, these procedures have only been tested in an Operations Manager 2007 R2 environment. Although Operations Manager 2007 R2 is mentioned throughout this guidance, the procedure should work the same on SP1 and RTM versions.
If you plan to use all or some of the computers that were previously used in the inoperable management group, we need to first ensure that all Operations Manager 2007 related components are removed. This will provide a clean slate for installing a new management group in phase 2.
Decommissioning the inoperable management group
Since we’ll be keeping the same computer names in this guidance, and assuming the management group name and computer names of core server roles will remain the same, we do not need to uninstall any agents. After installing the new management group, all agents should continue to report into the same management servers.
If you plan to rename the management group or computer names of core server roles, there will be additional configuration necessary on the agents. This is not the easiest recovery path, especially if under time constraints. These additional configuration steps are not outlined in this document.
If there is any issue uninstalling any of the above Operations Manager 2007 components from any computer, refer to the article reference in the link below for instructions on how to do a brute force uninstall. http://blogs.technet.com/b/jonathanalmquist/archive/2010/07/22/brute-force-uninstall-of-management-group.aspx
After phase 1 is complete, we can continue to phase 2 by installing a new management group.
Installing the new management group
This section is provided as a high-level overview of installation steps. It is not intended to replace the installation procedures outlined in Microsoft official documentation. To understand requirements, supported configuration and details installation steps, please refer to the official documentation referenced in the link below.
The default settings are to review new manual agent installations. In the case of a management group rebuild, the previously installed agents will be considered manually installed. You can either keep this setting and approve agents as they appear in Pending Management, or you can check the security setting option under Administration to automatically approve new manually installed agents. Checking this option will allow the previously agent-managed computers to automatically go into the Agent-Managed node without further administrative action.
If you chose to delete the data warehouse in this recovery procedure, you will need to install the data warehouse component on the computer hosting the SQL instance for the data warehouse before proceeding with the Report Server role installation.