Cisco Cisco TelePresence Management Suite (TMS) Version 15 Maintenance Manual
Database redundancy
Two of the SQL Server 2008 high-availability alternatives are appropriate for Cisco TMS: database mirroring
and failover clustering.
and failover clustering.
Database mirroring
Database mirroring increases database availability by maintaining a hot spare database on another SQL
Server instance. The primary SQL Server instance is referred to as the principal database. The principal
database has a secure channel to a mirror database residing on another SQL Server instance, preferably
located on another physical server.
Server instance. The primary SQL Server instance is referred to as the principal database. The principal
database has a secure channel to a mirror database residing on another SQL Server instance, preferably
located on another physical server.
In the database mirroring model, there are two types of transfer mechanisms, synchronous and
asynchronous.
asynchronous.
Synchronous transfer
If you are using synchronous mirroring, a transaction is not considered successfully completed until it is
committed in both the principal and mirror databases. This guarantees that the two databases are always
consistent. However, network speed and distance will affect database response times.
committed in both the principal and mirror databases. This guarantees that the two databases are always
consistent. However, network speed and distance will affect database response times.
Asynchronous transfer
If performance is more important than data consistency, data transfers can be set up as asynchronous
operations, but some data will then be lost in the case of a catastrophic failure in the principal database
server.
operations, but some data will then be lost in the case of a catastrophic failure in the principal database
server.
Network requirements for asynchronous mirroring:
Minimum bandwidth = 10Mbps
Latency < 50ms
Jitter < 1ms
Packet drops < 1%
Changing the database connection string
The SQL Server database mirroring feature supports automatic failovers between the principal and mirror
databases in the case of a failure on the principal database. However Cisco TMS does not,so the database
configuration must be manually changed in Cisco TMS. If you are using the NLB redundancy model, this
must be done on both Cisco TMS nodes:
databases in the case of a failure on the principal database. However Cisco TMS does not,so the database
configuration must be manually changed in Cisco TMS. If you are using the NLB redundancy model, this
must be done on both Cisco TMS nodes:
1. Go to the TMS Tools application
> Configuration > Change Database Connection Settings
.
2. Change the Database Server\Instance to point to the mirror database and click OK.
3. Restart all the TMS services as well as IIS to complete the manual failover operation.
Placing the principal and mirror databases in two different physical locations while using synchronous
mirroring is not supported by Cisco TMS, as transactions take too long to complete, causing slow responses
in the application.
mirroring is not supported by Cisco TMS, as transactions take too long to complete, causing slow responses
in the application.
Cisco TMS Administrator Guide (14.3.2)
Page 311 of 342
Redundant deployments
Database redundancy