About Multi-data Centers
With the Multi-data Center (MDC) feature, you can join two CWMS systems to create a single MDC system. You must purchase a license for each CWMS data center in an MDC system.
The meeting client uses the Round Trip Time (RTT) to determine on which data center to start meeting. (This is an automated process and cannot be configured by the host or the administrator.)
Load Balancing is not configurable; it is automatic and built into the system. Any Load Balancer configured as separate machine is not supported.
Network requirements between data centers can be found in the "Network Requirements for Multi-data Center" chapter of the CWMS Planning Guide at http://www.cisco.com/c/en/us/support/conferencing/webex-meetings-server/products-installation-and-configuration-guides-list.html.
Advantages of a Multi-data Center System
The advantages include:
End user access to all data centers by using one URL and one set of phone numbers; the existence of MDC is transparent to end users.
Host licenses, recordings, and related management data migrate freely between joined data centers.
Users can dial into meetings without geographic restrictions; attend meetings by dialing local phone numbers.
Data centers can (optionally) be located in different geographic areas.
Zero-downtime during some planned maintenance events, when the data centers can be running different CWMS 2.5 update versions. Consult the release notes at http://www.cisco.com/c/en/us/support/conferencing/webex-meetings-server/products-release-notes-list.html to determine which CWMS versions can run simultaneously.
Occasionally, data centers in an MDC system can be running different update versions. Consult the release notes at http://www.cisco.com/c/en/us/support/conferencing/webex-meetings-server/products-release-notes-list.html to determine which CWMS versions can run simultaneously.
A disaster recovery environment that is transparent to users. If one data center fails for any reason, the other data center supports users.
Although in an MDC environment the data centers are all running CWMS and considered peers, for the purpose of describing the process for joining data centers in a system, the relationship between data centers are considered primary and secondary. Before the Join, the primary data center supports the system you want to retain, and shall be the location of the license manager. The secondary data center becomes part of the MDC system. The distinction is important especially if you are joining data centers that have been actively supporting users. User information and content are deleted from the secondary data center.
There is redundancy, but no increase in capacity when a data center is added to an MDC system. If a 2000-port data center is added to an MDC system supported by a 2000-port data center the resulting system is a 2000-port MDC.
If you are joining a new, secondary CWMS system data center that has no user data to an MDC system, continue to Preparing an MDC System to Receive Data Center Join Requests.
If you are joining an active, secondary CWMS system data center that includes user data to an MDC system, continue to Preparing to Join an Active CWMS Data Center to a MDC System.
Each data center polls its components for their status every 5 seconds. Under certain failure conditions a data center automatically turns on Blocking Mode to prevent end users from accessing a data center with failed components, allowing time for the system to attempt to fix itself. In an MDC environment, user activity transparently fails over to the active data center. Once the components of the data center in Blocking Mode are again operational, the data center exits Blocking Mode. Email notifications are sent to administrators when a data center goes into or recovers from Blocking Mode.
Blocking Mode ON conditions are triggered when all the following are true:
One or more of the telephony components or data base replication fails.
The condition has existed for 5 minutes or more.
Another data center in the MDC system is operational.
End-user access to a data center in Blocking Mode is prevented; all user activity is redirected to the active data center. Administrators can access the administration site on the blocked data center to monitor its condition and to troubleshoot issues.
Blocking Mode OFF conditions are automatic and are triggered by all of the components returning to a good state. Access by end-users is restored and the data center returns to polling its components every 5 seconds.