Some situations require that you remove all server roles, Standard Edition servers, and Front End pools at all sites, and that you roll back the environment. For example, in a disaster recovery scenario, a hardware failure scenario, an upgrade scenario, or other situation where you must completely remove the servers, pools, and settings.
Before you decommission a deployment, do the following:
Provide users with sufficient notice to ensure that they can plan for the down time.
Deactivate the server roles on the weekend or holiday, or during other off-peak hours.
Disable monitoring in System Center Operations Manager or third-party monitoring system
Delete all Contact objects enabled for Lync Server Enterprise Voice features by using Lync Server Management Shell.
If the Lync Server 2010 response groups have been migrated to a Lync Server 2013 deployment, do not remove the contact objects of the migrated response groups. If response groups have been migrated, skip the document “Remove Response Group Service Workflow Contact Objects” that is listed in the “Documentation” column. For more information, see http://technet.microsoft.com/en-us/library/jj204854.aspx.
Remove Enterprise Voice routes by using Lync Server Control Panel.
Remove all Call Park orbits by using Lync Server Control Panel.
Remove all tables for Enterprise Voice unassigned phone numbers.
Back up the Location Information service database.
Back up the custom music on hold file.
Delete Enterprise Voice routes.
Reassign the public switched telephone network (PSTN) gateway.
Phase 13: (Optional) Check for existence of the Active Directory Domain Services (AD DS) service control point (SCP) for the Central Management store, and if it is still defined, remove it by using Lync Server Management Shell.
This is only necessary if you are decommissioning the deployment to install a new deployment in which the Central Management store is hosted by a different SQL Server-based Back End Server.