Maintenance Manual

2-15
Cisco ICM Enterprise Edition Administrator Guide Release 6.0(0)
Chapter 2 Fault Tolerance
Database Fault Tolerance
After SQL Server automatic recovery is completed, the off-line Logger
synchronizes its state with the state of the on-line Logger. After the state transfer
process takes place, both members of the Logger pair can execute as a
synchronized process pair.
During the time that one database is off-line, configuration data may have been
added to the contents of the on-line database. If any configuration data changed
while one database was off-line, the configuration changes are applied to the
database as part of the Logger’s state transfer process. This configuration update
happens as part of the state transfer before synchronized execution begins.
Any historical data that accumulated in the on-line database is recovered after
synchronized execution begins. Rather than attempting to recover the historical
data immediately, ICM software first restores system fault tolerance (that is,
duplexed database capability and synchronized execution).
ICM software recovers historical data from the on-line database using a special
process called Recovery. In Recovery, the Logger communicates with its
counterpart on the other side of the Central Controller and requests any historical
data that was inserted during the off-line period. The counterpart delivers the data
over the private network that connects both sides of a duplexed Central Controller.
Disk Failure
A disk failure requires additional steps. If a disk failure disables one side of the
Central Controller database, the disk must be repaired or replaced.
Note Contact your ICM support provider if a disk failure occurs.
Your support provider may repair or replace the disk and perform the following
steps:
Step 1 Rebuild the database structure from scratch on the new disks.
Step 2 Restore the configuration data, either from:
• A snapshot of the on-line database.
• The most recent backup tape.