[This is pre-release documentation and subject to change in future releases. This topic's current status is: Milestone-Ready]

Topic Last Modified: 2010-07-18

Microsoft Communications Server 2010 introduces a new server draining feature that enables you to take a server offline (for example, to apply software or hardware upgrades) without any loss of service to users. When you drain one server in a pool, it stops taking new connections and calls. These new connections and calls are routed through other servers in the pool. A server being drained allows its sessions on existing connections to continue until they naturally end. When all existing sessions have ended, the server is ready to be taken offline.

When a Front End Server is drained, some Communications Server 2010 features and services rely on the new DNS load balancing feature to ensure proper draining. If you are not using DNS load balancing on the pool, connections through these services may not be re-routed to other servers during the draining process, and thus when the server is taken offline some sessions and calls may be interrupted. The features that rely on DNS load balancing to ensure draining are as follows.

For more information about DNS load balancing, see DNS Load Balancing.

In addition to draining all services on a server running Communications Server 2010, you can also drain individual Communications Server services, for instances when you are applying a Communications Server patch that does not require the whole server to be shut down. Note that when you drain one service, you must select a service as it is grouped and displayed in the Windows list of services. For example, the Communications Server Front End service and the data collection agent for Monitoring Server are separate Communications Server services, but in the Windows services list they are consolidated and shown as the RTCSrv.exe service. You can drain the RTCSrv.exe service, but you cannot drain individually the two underlying Communications Server services.