Info2soft use cookies to help you have a superior and more admissible browsing experience on our website. Privacy Policy
Oracle Hot Backup Mode lets administrators copy datafiles while the database stays open. It is used for user-managed online backups when taking the database offline is not an option.
Oracle continues processing transactions during this mode, but the corresponding redo logs need to be preserved and applied to make the copied files consistent.
Oracle Hot Backup Mode is a database state that allows datafiles to be copied while the database remains open and active. It is designed for user-managed online backups, so backup operations can run without interrupting normal transactions.
When a tablespace or the entire database enters backup mode, Oracle freezes checkpoint updates in the relevant datafile headers. Reads and writes continue as normal throughout.
Key characteristics:
Unlike a cold backup, hot backup mode does not require a database shutdown. It does not, however, guarantee that copied files are immediately recoverable without the required redo logs.
Hot Backup Mode relies on two mechanisms to support online backups: checkpoint SCN freezing and redo-based recovery.
1. Checkpoint SCNs and fractured blocks: When a tablespace enters backup mode, Oracle freezes checkpoint updates to the headers of its associated datafiles.
A checkpoint SCN (System Change Number) tracks which changes are already reflected in a datafile — freezing it tells Oracle where recovery needs to begin. Database writes continue normally, which creates a risk called a fractured block, where a copied block contains parts from different points in time.
To handle this, Oracle records full block images in the redo logs for blocks modified during backup mode, alongside normal redo records.
2. Redo generation and archived redo logs: Because Oracle writes additional full block images during backup mode, redo generation increases. Archive destinations should have sufficient space, and backup mode should stay active only for as long as it takes to copy the files.
Copied datafiles are not a consistent backup on their own — Oracle needs the corresponding archived redo logs to recover them to a consistent state. These logs should be retained alongside the datafile copies.
Oracle supports two approaches for entering hot backup mode: at the tablespace level for selected datafiles, or at the database level for all datafiles at once. The right choice depends on the scope of the backup.
Traditional user-managed online backups require the database to run in ARCHIVELOG mode, which allows Oracle to archive the redo logs needed to recover copied datafiles.
Check the current archiving mode with:
ARCHIVE LOG LIST;
The output should show:
Database log mode Archive Mode
Automatic archival Enabled
If the database is in NOARCHIVELOG mode, it needs to be switched to ARCHIVELOG mode before proceeding.
Before starting, also confirm:
To place a specific tablespace into backup mode:
ALTER TABLESPACE users BEGIN BACKUP;
This freezes checkpoint updates to the headers of the tablespace’s datafiles while normal operations continue. Before copying, identify the associated datafiles:
SELECT FILE_NAME
FROM DBA_DATA_FILES
WHERE TABLESPACE_NAME = 'USERS';
Replace USERS with the actual tablespace name. The query returns the file paths that need to be copied.
To place all datafiles into backup mode at once:
ALTER DATABASE BEGIN BACKUP;
This is useful when a coordinated full database file copy is needed. Keep in mind that all tablespaces enter backup mode simultaneously, which increases overall redo generation for the duration of the operation.
Backup mode should be ended as soon as the required datafiles have been copied.
For a tablespace-level backup:
ALTER TABLESPACE users END BACKUP;
For a database-level backup:
ALTER DATABASE END BACKUP;
After ending backup mode, archive the current redo log using ALTER SYSTEM ARCHIVE LOG CURRENT to ensure the redo needed for recovery is preserved.
ALTER SYSTEM ARCHIVE LOG CURRENT;
END BACKUP does not make the copied datafiles immediately recoverable on its own. The corresponding archived redo logs, control file information, and media recovery are still required.The following example demonstrates a traditional user-managed backup of the USERS tablespace on a Linux system.
Query DBA_DATA_FILES to get the file paths associated with the target tablespace:
SELECT FILE_NAME
FROM DBA_DATA_FILES
WHERE TABLESPACE_NAME = 'USERS';
Record the returned paths and confirm the backup destination has enough space. A complete backup strategy should also account for control files and relevant database metadata.
Enable backup mode for the USERS tablespace:
ALTER TABLESPACE users BEGIN BACKUP;
Oracle freezes checkpoint updates to the tablespace’s datafile headers while transactions continue normally. Confirm backup mode is active before copying any files:
SELECT FILE#, STATUS
FROM V$BACKUP
WHERE STATUS = 'ACTIVE';
Use an OS utility to copy the datafiles to a separate backup location. On Linux:
cp /u01/oradata/ORCL/users01.dbf /backup/
Keep the copy window as short as possible to limit additional redo generation. The copied files are not yet recoverable at this stage — the corresponding redo logs are also required.
End backup mode as soon as the files have been copied:
ALTER TABLESPACE users END BACKUP;
Then archive the current redo log using ALTER SYSTEM ARCHIVE LOG CURRENT:
ALTER SYSTEM ARCHIVE LOG CURRENT;
The archived redo logs needed for recovery should be stored alongside the datafile copies.
Check that no datafiles remain in backup mode:
SELECT FILE#, STATUS
FROM V$BACKUP
WHERE STATUS = 'ACTIVE';
An empty result means no files are still in active backup mode. Also confirm:
A file copy alone does not guarantee recoverability. The backup should be tested by restoring and running media recovery in a separate environment.
The commands above illustrate a basic tablespace-level workflow and are not a complete production backup script. A production backup also requires handling control files, archived redo logs, and recovery validation.
Incorrect backup procedures can cause operational and recovery problems. The following covers common issues encountered during traditional user-managed online backups.
A failed script, interrupted file copy, or forgotten END BACKUP command can leave datafiles in backup mode longer than intended, causing unnecessary redo generation.
Query V$BACKUP to identify any affected files:
SELECT FILE#, STATUS
FROM V$BACKUP
WHERE STATUS = 'ACTIVE';
If active files are found, end backup mode with the appropriate command:
ALTER TABLESPACE users END BACKUP;
-- or
ALTER DATABASE END BACKUP;
If the database crashed or the backup was interrupted mid-operation, check the database state and follow Oracle’s recovery procedure before making further changes.
Hot backup mode generates additional full block images for modified blocks, so heavy write activity or a long backup window can increase redo volume and pressure on archive destinations.
To keep this manageable:
Copying a datafile outside the correct backup mode procedure, or interrupting the copy mid-way, can produce a file that cannot be recovered consistently.
If a backup looks incomplete or suspect:
If the required recovery information is missing, a new backup is needed.
Online backups depend on archived redo logs to bring copied datafiles to a consistent state. Missing logs can make recovery impossible regardless of whether the datafile copies are intact.
BEGIN BACKUP through END BACKUP for every backup set.User-managed online backups require careful coordination of backup mode commands, file copies, archived redo logs, and recovery validation. For teams managing multiple Oracle instances or mixed database environments, this overhead adds up quickly.
i2Backup is an enterprise backup solution that handles Oracle database protection without the manual steps involved in traditional hot backup workflows.
BEGIN BACKUP and END BACKUP commands manually.For teams looking to move beyond manual Oracle backup procedures, i2Backup offers a more automated and consistent approach to Oracle database backup and recovery.
Evaluate Your
Backup and Recovery Capabilities
Assess backup and recovery capabilities to find a solution that meets your organization’s data protection needs.
Q1: Does Oracle Hot Backup Mode Affect Database Availability?
No. The database stays open and accessible throughout the backup. Users can continue reading and writing data as normal.
Q2: Does Hot Backup Mode Freeze Datafile Writes?
No. It freezes checkpoint updates to datafile headers only. Database writes continue normally during the backup.
Q3: Can You Enable Hot Backup Mode in NOARCHIVELOG Mode?
No. Traditional user-managed online backups require ARCHIVELOG mode. Oracle needs archived redo logs to recover the copied datafiles after the backup.
Q4: How Can You Check Whether a Tablespace Is in Backup Mode?
Query V$BACKUP:
SELECT FILE#, STATUS FROM V$BACKUP;
ACTIVE means the datafile is in backup mode. NOT ACTIVE means it is not.
Q5: What Happens If You Forget to Run END BACKUP?
The datafiles stay in backup mode and Oracle continues generating extra redo. If the database crashes while files are still active, media recovery will be required at the next startup. Run ALTER DATABASE END BACKUP; to clear the state.
Q6: Is Oracle Hot Backup Mode Deprecated?
No. User-managed hot backup is still supported in current Oracle versions. However, Oracle recommends RMAN for most backup workflows, as it handles consistency and recovery automatically without manual BEGIN BACKUP and END BACKUP commands.
Oracle Hot Backup Mode lets you copy datafiles without taking the database offline, but it requires careful execution. End backup mode promptly after copying, and retain the corresponding archived redo logs for recovery to succeed.
For most production environments, RMAN is the more practical choice — it handles backup consistency and recovery management without manual commands. For teams looking for a more automated approach to database protection, Info2soft offers solutions covering backup, recovery, and replication across Oracle and other database platforms.
· Enterprise & Mid-market Customers Worldwide.
· Support team available to assist you throughout your trial.
· Start a 60-day free trial or view demo to see how Info2soft protects enterprise data.