Maintenance Manual

Chapter 2 Fault Tolerance
Database Fault Tolerance
2-14
Cisco ICM Enterprise Edition Administrator Guide Release 6.0(0)
available). Therefore, when a CallRouter fails and restarts, at least one Logger and
central database must be available so that the CallRouter can load the
configuration data into memory.
In addition to configuration data, Peripheral Gateways, NICs, and the CallRouter
itself all produce historical data. The system components gather historical data
and pass it to the CallRouter, which then delivers it to the Logger and the central
database. The Logger passes the historical data on to an Historical Data Server
(HDS) facility on a Distributor Admin Workstation.
The ability of the CallRouter to deliver data to the Logger and the central database
is not necessary for call routing. However, the ICM’s monitoring and reporting
facilities require both real-time data and historical data from the central database.
Database fault tolerance and data recovery, therefore, are extremely important to
the reporting functions of ICM software.
ICM Database Recovery
Database recovery is the process of bringing an off-line database up to the same
state as an on-line database. In a database device failure (for example, in a disk
failure), some manual intervention is required to restore duplexed operation and
bring the off-line database up to date. The following scenarios describe what
happens in a system failure, a disk failure, and a software failure.
System Failure
When a single Logger, CallRouter, or Database Manager fails (for example, due
to a power outage), the associated central database will go off-line. The process
of bringing the off-line database back to full synchronization is completely
automatic. If the Logger machine reboots, SQL Server automatic recovery runs to
ensure that the database is consistent and that all transactions committed before
the failure are recorded on disk.
Note If the Logger machine does not reboot, SQL Server automatic recovery is not
required.