Hi There,

I found slow performing pl/sql functionality indicated to be slow only when a backup of rachivelog is in progress.

Checking the events using AWR and ASH reports there is clear indication for "Log file sync" waiting event for a delete from temporary table. As this is always happening while the archivelog backup is running I just decided to check the backup script.

And here it is:

run{
SQL 'ALTER SYSTEM ARCHIVE LOG CURRENT';
backup current controlfile;
SQL 'ALTER SYSTEM ARCHIVE LOG CURRENT';
backup archivelog all not backed up 1 times;
DELETE archivelog UNTIL TIME 'SYSDATE-1' BACKED UP 1 TIMES TO DEVICE TYPE DISK;
}


I do not see a valid reason why this backup have "ALTER SYSTEM ARCHIVE LOG CURRENT" 3 times in the backup script.
As I am aware the "ARCHIVE LOG CURRENT" is automatically initiated when "BACKUP ARCHIVELOG ALL" is used.

Also I am while the checkpoint is in progress no one can write to the redo logs and this is happening every 1h. on this database for really long time.

Could you please provide with your opinion on this backup setup and advice for the best way to backup the archives on hourly base.

My option is:

run{
backup archivelog all not backed up 1 times;
DELETE archivelog UNTIL TIME 'SYSDATE-1' BACKED UP 1 TIMES TO DEVICE TYPE DISK;
backup current controlfile;
}