TANDBERG TMS SERVER REDUNDANCY - CONFIGURATION GUIDE 13.0 Configuration Manual - page 23
Cisco Redundancy Models for Cisco TMS
Cisco TMS redundancy configuration and overview
Page 23 of 32
installation completes and the server reboots if necessary. Repeat these steps one server at a
time until all Cisco TMS servers have been upgraded
Once all Cisco TMS servers have been upgraded, you may restart all Cisco TMS services and
WWWPublishingService on all Cisco TMS servers.
Log into the Cisco TMS website, go to Administrative Tools>Configuration>Network and
change the Cisco TMS Server IPv4 Address (Local): address to the IP address that is forwarded
by the network load balancer. Your upgrade is complete
Upgrades will erase any changes to the IIS web configuration files, including the viewstate
machine key and method defined during installation. Repeat the viewstate key generation and
installation steps outlined in the Installation section on each of the upgraded web servers
Model Summary
Immediate fail-over available.
Yes for SQL Server. Yes for Cisco TMS Server
Amount of data lost.
No data will be lost if fail-over has occurred, as Cisco TMS and the SQL Server can failover
automatically. Please be sure to read Section
0 Cisco TMS Redundancy limitations and notes
for
specific notes about processes during a server failure.
Time to restore service.
For SQL Server, recovery is dependant on SQL Server configuration, normally on the order of
seconds to minutes. For Cisco TMS Server, dependant on NLB configuration but should be
transparent to the users
Administrative Capabilities Required.
High for Network Load Balancer and handling multiple Cisco TMS servers. Very high for SQL
Cluster.
Cost.
Highest. Must implement multiple servers, Network Load Balancer, and SQL Cluster if it does not
already exist
Manual Off-Site Fail-Over
While using a network load balancer and SQL cluster will provide automatic fail-over for both the Cisco
TMS server and SQL database, the solution requires all the servers be located together. Some
customers require a fail-over solution that allows cut-over to an installation off-site from the primary
installation in case of catastrophic network failure at the primary site. The basis of an off-site fail-over
is to maintain a copy of the Cisco TMS Server and Cisco TMS database that are inactive in an off-site
location. When fail-over is required, an administrator cuts over to the secondary location. One solution
for this is to use Log Shipping, a functionality of Microsoft SQL Server to keep the database in the
secondary location up to date. Log Shipping allows you to have one main database, and a separate
fail-over database. Log Shipping allows all changes to the main database to be batch transferred to
the fail-over database off-site. This model uses a complete ‘copy’ of the primary system off-site, so this
technology can be combined with other redundancy models within each location if desired. It is critical
that when combining this model with other
methods, such as using a Network Load Balancer
at each location, that the off-site location must be
completely disabled; all Cisco TMS services and
websites, until the off-site location is activated to
become the primary site.
Microsoft references on SQL Server Log Shipping
http://www.microsoft.com/technet/prodtechnol/sql/
2000/deploy/hasog02.mspx
Log shipping to a secondary site provides high
resiliency to network failure, but has some
disadvantages as well.
Error! Objects cannot be created from editing field codes.
Figure 3 Balancing Off-Site Cutover