When a database instance is restarted, the database system switches it from operational state ADMIN or ONLINE to the OFFLINE operational state, and then switches it to the ONLINE operational state again.

A restart is necessary, for example, in the following situations:

?     You have changed a database parameter and the change can only take effect after a restart.

?     The database instance was not stopped correctly because the computer was switched off due to a power failure.

For a restart, the database system uses information from the last savepoint (most recent consistent status of the database instance stored in the volumes), and the log entries still stored in the volumes, in order to bring the database instance as close to the state it was in when it was switched to the OFFLINE operational state. 

For a restart to be successful, the volumes in the database instance must not be damaged. If the hard disk has been damaged, you must restore the database instance.

You can also restart the database instance in a state it was in at a point in time in the past.

Each time the database instance is started, a copy of the kernel log is created under that name, and the old kernel log is overwritten with new information.

See also:

Concepts of the Database System, Example: Restart, Savepoint

Database Manager GUI, Starting (Restarting) the Database Instance

Database Manager CLI, db_restart