Alter System ‘SWITCH LOGFILE’ vs ‘ARCHIVE LOG CURRENT’

Both the command is used to force the log switch but they do it in different way. ARCHIVELOG CURRENT waits for the writing to complete. This can take several minute to complete for large redologs file where as SWITCH LOGFILE is fast as it does not wait for the archiver process to complete writing the online redologs.

The ‘ARCHIVELOG CURRENT’ command is safer because it waits for OS to acknowledge that the redolog has been successfully written. Hence this command is best practice for production backup script for RMAN.

The ‘ARCHIVELOG CURRENT’ allows, you to specify the thread to archive (If you do not pass the thread parameter then oracle will archive all full online redo logs) where as ‘SWITCH LOGFILE’ archives only the current thread.

Note: In case of RAC environment, the ‘ARCHIVELOG CURRENT’ will switch the log of all nodes (instances) where as ‘SWITCH LOGFILE’ will only switch the log file on the instance where you issue this command.

SWITCH LOGFILE’ is not a correct method after hot backup

In order to be sure that all changes generated during a hot backup are available for a recovery, Oracle recommends that the current online log be archived after the last tablespace in a hot backup is taken out of backup mode. A similar recommendation stands for after a recovery manager online database backup is taken.

Many DBA’s or Oracle books will suggest using the ‘ALTER SYSTEM SWITCH LOGFILE’ command to achieve this. The use of ‘SWITCH LOGFILE’ to obtain archives of the current logs is not a perfect to follow because the command returns success as soon as the log writer has moved to the next log, but before the previous log has been completely archived. This could allow a backup script to begin copying that archived redolog before it is completely archived, resulting in an incomplete copy of the log in the backup.

Thus the better command to use for archiving the current log is ‘ALTER SYSTEM ARCHIVE LOG CURRENT’ This command will not return until the current log is completely archived. For the same reasons outlined above, a backup script should never just back up “all the archived redologs”. If a script does not restrict itself only to those logs that it knows to be archived, it may improperly try to back up an archived redolog that is not yet completely archived.

To determine which logs are archived, a backup script should query the v$archived_log view to obtain a file list for copying.

Advertisements
Leave a comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: