The replication between two Symantec Endpoint Protection Manager (SEPM) sites fails after several weeks due to Microsoft SQL query timeouts.

This can happen when the indexes of the Microsoft SQL database used by the Symantec Endpoint Protection Manager (SEPM) in one or both sites become significantly fragmented. When this happens, the Microsoft SQL server is not able to use these indexes when running the queries used by the manager to generate replication data. This causes the queries to take exponentially longer than queries run against a properly indexed database.

To work around this problem, rebuild the indexes on the following database tables: SEM_AGENT, SEM_CLIENT, and SEM_COMPUTER.

For more information on rebuilding Microsoft SQL indexes, see http://technet.microsoft.com/en-us/library/ms189858.aspx.