The role of the recovery log

When the logs that make up the recovery log are set up carefully, they work together to ensure that data is not lost.

Each component of the recovery log has a specific purpose to protect server and client data:
Active log
The active log records transactions that are in progress on the server. This information is needed to restart the server and database after a disaster. Transactions are stored in the log files of the active log, and a transaction can span multiple log files. If a failure occurs, the changes that were made but not committed are rolled back, and all committed transactions, which might not have been physically written to disk, are reapplied and committed again.
Active log mirror (optional)
The active log mirror is a copy of the active log that can be used if the active log files cannot be read. All changes that are made to the active log are also written to the log mirror. There can be only one active log mirror. Creating a log mirror is optional but recommended.
Archive log
The archive log contains copies of closed log files that were in the active log. The archive log is not needed for normal processing, but it is typically needed for recovery of the database.

The archive log is included in database backups and is used for rollforward recovery of the database. Archive log files that are included in a database backup are automatically pruned after a full database backup cycle is complete. The archive log must have enough space to store the log files for database backups.

Archive failover log (optional)
The archive failover log, also called a secondary archive log, is the directory that the server uses to store archive log files when the archive log directory is full. Its use is optional but highly recommended.