Online redo log group archiving is performed by either the Archive (ARCH) background process when automatic archiving is started or by a user process that issues SQL statements to archive the online redo log groups manually. The ARCH process archives a redo log group after the group becomes inactive and the log switch to the next online redo log group has completed.
At this point the record is created in the databases control file identifying the archived redo log file. The group of online redo log files being archived cannot be reused and written to by the LGWR process until the ARCH process has concluded and released a lock on the redo log files. This ARCH locking process guarentees that the LGWR process does not accidently overwrite online redo log file that needs to be archived. When the ARCH process has completed archiving the redo log files, a second record is written to the databases control file identifying the success of the archiving process.
The archive process uses the LOG_ARCHIVE_DEST parameter to specify the destination of the archived redo log files. This destination is usually a storage device separete from the Oracle database. The ARCH process can write out two copies of an online redo log group - the duplexing feature . The initialization parameter LOG_ARCHIVE_DUPLEX_n specifies the location of the archived redo log file. The parameter LOG_ARCHIVE_MIN_SUCCEED_DEST specifies the number of archive log destinations to which redo log group must be successfully arhcived. A missing archived redo log file renders all subsequent archived redo log files usless.
To multiplex the controlfiles:
CONTROL_FILESand indicate where the new mirrored (multiplexed) control-file can can be accessed.
To Multiplex redo log files :
The multiplexing is specified while the database is in mount and is done through the
alter database command , the following creates a new log group.
alter database >add logfile group 5 > ('/u02/app/oradata/OSMOSIS/log_1c.dbf', > /u05/app/oradata/OSMOSIS/log_2c.dbf') size 5M;
To add a new log file to an existing group:
alter database >add logfile member '/u05/app/oradata/OSMOSIS/log_3c.dbf' >to group 3;
When adding a file to a redo log group there is no need to specify the size, because the group has already been indicated , besides all the redo log files in the group must be of the same size.
The following are types of failures: