Oracle AI Database 26ai RMAN is a course in using Oracle Recovery Manager (RMAN) to back up and recover Oracle databases on the CDB and PDB architecture, including automated scripts, complete and point-in-time recovery, and backups to disk, tape and the cloud. It suits database administrators and system engineers who already look after Oracle Database.
The data in an Oracle database is often the heart of an organisation's systems, yet many sites still have backups that have never been test-restored, do not know whether they can roll data back to before a mistake, and have no clear procedure for when a disk fails or a whole server is lost. Oracle Recovery Manager (RMAN) is the main tool for this work. It works directly with the database, detects block-level corruption during backup and restore, and turns backup and recovery into a planned, repeatable task.
This course is updated for Oracle AI Database 26ai, the current release that succeeds 23ai and uses the CDB and PDB architecture. It starts with a strategy based on RPO and RTO, setting up ARCHIVELOG mode, the fast recovery area and RMAN, backing up whole databases and parts of them, using a recovery catalog and writing scripts that run automatically. Learners then practise complete recovery, point-in-time recovery, Flashback and repairing corrupt blocks, before moving on to backups to disk, tape and the cloud, restoring to new disks or a new server, and a closing workshop that writes a backup plan and rehearses a full system recovery. (3 days, 6 hours per day, 18 hours in total, Intermediate level.)
What you’ll gain
Plan a backup and recovery strategy around the organisation's RPO and RTO
Configure RMAN, ARCHIVELOG mode and the fast recovery area on CDBs and PDBs
Back up whole databases and parts of them with full and incremental backups
Write RMAN scripts and schedule them to run automatically
Perform complete and point-in-time recovery, including at PDB and table level
Back up to and restore from disk, tape and cloud object storage
Restore a database to new disks or a new server when the original hardware fails
Who this course is for
Database administrators (DBAs) who look after Oracle Database in their organisation
System engineers responsible for backup and system recovery
People responsible for backing up and recovering Oracle databases who want a systematic plan
Teams moving from older Oracle releases to 19c, 23ai or 26ai
Prerequisites
Prior knowledge of Oracle Database administration
An understanding of Oracle files such as datafiles, control files and redo logs
Basic Linux commands, such as managing files, editing configuration files and using sudo
A laptop with at least 16 GB of RAM that can run a VM, with permission to install software
Curriculum
Course Details
This course is the 2026 edition of Oracle Database 21c RMAN, updated for Oracle AI Database 26ai, the current release that succeeds 23ai and is built on the CDB and PDB architecture. It runs for 3 days, 6 hours per day (18 hours in total, 09:00-16:00), as lectures with labs on a database VM that learners use throughout. Intermediate level. It covers strategy, configuration, full and incremental backups, complete and point-in-time recovery, Flashback, backups to tape and the cloud, and restoring to a new server, and most of the content also applies to Oracle Database 19c. Learners take home a lab guide, RMAN and shell scripts, a backup policy template and a recovery runbook.
Day 1Backup Principles and Configuring RMAN
Section 1: Lab: Backup Strategy and the 26ai Environment
Types of failure, and setting the organisation's RPO and RTO
Oracle AI Database 26ai and the CDB and PDB architecture as the only model
Files to protect: datafiles, control files, redo logs, archived logs and the SPFILE
Enable ARCHIVELOG mode and set up the fast recovery area
Lab: prepare the database VM used throughout the course
Section 2: RMAN Architecture and Backup Types
The RMAN client, target database, channels and repository
Keeping backup metadata in the control file versus a recovery catalog
Backup sets versus image copies, and consistent versus inconsistent backups
Whole-database backups versus partial backups
Lab: connect RMAN to the CDB and a PDB and review the defaults
Section 3: Lab: Configuring RMAN Parameters
SHOW ALL and the CONFIGURE settings to make on day one
Redundancy and recovery window retention policies
Control file autobackup, channels and file name formats
Compression, and backup encryption, which defaults to AES-256 in newer releases
Lab: configure RMAN to match a sample backup policy
Section 4: Lab: Backing Up Whole Databases and Parts of Them
BACKUP DATABASE PLUS ARCHIVELOG at CDB level
Back up a single PDB, tablespace or datafile
Level 0 and level 1 incremental backups, differential and cumulative
Block change tracking to speed up incremental backups
Lab: build a weekly schedule of full and incremental backups
Section 5: Recovery Catalog and Automated RMAN Scripts
Create a recovery catalog, register databases and use stored scripts
Write RMAN scripts on Linux and Windows to automate tasks
Schedule them with cron, Task Scheduler or DBMS_SCHEDULER, with logging
LIST, REPORT, CROSSCHECK and DELETE OBSOLETE
Lab: a backup script that sends an alert when it fails
Day 2Recovering the Database in Different Situations
Section 6: Restore and Recovery Concepts
Instance failure, media failure and user error
How restore differs from recover, and the role of redo
Complete versus incomplete recovery
Check readiness with RESTORE ... PREVIEW and RESTORE ... VALIDATE
A manual diagnosis workflow in place of the desupported Data Recovery Advisor
Section 7: Lab: Complete Recovery
A regular datafile lost while the database is open
A damaged SYSTEM or UNDO datafile
Recover a single PDB while other PDBs keep running
A lost control file restored from autobackup
A damaged redo log member or group
Section 8: Lab: Point-in-Time Recovery
SET UNTIL TIME, SCN and SEQUENCE
Database point-in-time recovery, OPEN RESETLOGS and incarnations
PDB point-in-time recovery and tablespace point-in-time recovery
RECOVER TABLE to restore just a table that was dropped or changed by mistake
Lab: roll data back to before an accidental delete
Section 9: Flashback Technologies
Flashback Database and guaranteed restore points
Flashback Table, Flashback Drop and Flashback Query
Using a restore point before an upgrade or major change
Choosing Flashback or RMAN for the situation
Lab: rewind the database with a restore point
Section 10: Detecting and Repairing Corrupt Blocks
Physical and logical corruption, and block checking settings
BACKUP VALIDATE and VALIDATE DATABASE, PDB or datafile
Block media recovery with RECOVER ... BLOCK
Regularly checking that existing backups can really be restored
Lab: simulate a corrupt block and repair it without shutting down
Day 3Tape, Cloud, New Servers and a Recovery Plan
Section 11: Backing Up to and Restoring from Disk and Tape
DISK and SBT channels and media management software
BACKUP RECOVERY AREA and backing up your backups
Multisection backups for very large files
Retention and media rotation between disk and tape
Lab: simulate an SBT destination, then back up and restore
Section 12: Cloud Backups and Immutable Backups
The Oracle Database Cloud Backup Module with OCI Object Storage
Buckets with retention rules to stop backups being deleted or changed
Encrypting backups and keeping keys or wallets safe
An overview of Autonomous Recovery Service and Recovery Appliance
Facing ransomware with separate copies and a recovery plan
Section 13: Lab: Restoring to New Disks or a New Server
SET NEWNAME and SWITCH to restore onto new disks when the old ones fail
Restore a whole database to a new server
DUPLICATE to create a database for testing
Clone a PDB from backup for development work
Lab: recover the system onto a new server from the latest backup
Section 14: Tuning and Monitoring Backup Jobs
Monitor jobs with V$RMAN_BACKUP_JOB_DETAILS and V$SESSION_LONGOPS
Parallelism, multisection backups and compression levels
Find bottlenecks between disk, network and CPU
Backup status reports that managers can understand
Lab: compare backup times as settings change
Section 15: Workshop: Backup Plan and Recovery Drill
Write a backup policy for a sample system based on its RPO and RTO
Create a runbook of recovery steps for each scenario
Rehearse recovery from an unannounced simulated incident
An overview of Data Guard for systems that need a standby
A monthly checklist for the health of your backups
Schedule & training options
For individuals — public rounds
No public rounds are open right now. Join the waiting list and we will contact you first when the next round opens, or ask us on LINE. Or call 02-570-8449 or 088-807-9770
What is RMAN, and how does it differ from backing up an Oracle database by copying files yourself?
RMAN (Recovery Manager) is the backup and recovery tool that ships with Oracle Database and works directly with it. It can back up an open database, check blocks for corruption during the backup, take incremental backups, keep track of which backups exist and where, and pick the files needed for a restore automatically. Copying files yourself means handling backup mode, tracking files and checking integrity by hand, which is far more error-prone.
What is the difference between complete recovery and point-in-time recovery in Oracle?
Complete recovery restores the damaged files and applies all the redo so the data comes back to the present with nothing lost. It is used when files or disks are lost. Point-in-time recovery rolls data back to a chosen time, SCN or log sequence, and is used after data is deleted or changed by mistake, so anything after that point is lost. In 26ai it works at CDB, PDB and tablespace level, and RECOVER TABLE can recover a single table.
What are the steps to write an RMAN script that backs up a database automatically?
First set the standing RMAN settings with CONFIGURE, such as the retention policy and control file autobackup. Then write an RMAN command file, for example BACKUP DATABASE PLUS ARCHIVELOG followed by DELETE OBSOLETE, and call it from a shell script or batch file that sets the environment and keeps a log. Schedule it with cron, Task Scheduler or DBMS_SCHEDULER, and check the result in the log or V$RMAN_BACKUP_JOB_DETAILS, with an alert when it fails.
Who is the Oracle AI Database 26ai RMAN course for, and what should I know beforehand?
It suits DBAs, system engineers and anyone responsible for backing up and recovering Oracle databases, including teams moving from older Oracle releases. You should already know Oracle Database administration, understand Oracle's file structure and be able to use basic Linux commands, as the labs run on Oracle Linux in a VM.
What will I be able to do after the Oracle AI Database 26ai RMAN course?
You will be able to plan a backup strategy around RPO and RTO, configure RMAN, run full and incremental backups with automated scripts, perform complete and point-in-time recovery at CDB, PDB, tablespace and table level, use Flashback, repair corrupt blocks, back up to disk, tape and the cloud, and restore to new disks or a new server when hardware fails. Participants who complete the training receive a Certificate of Completion with its own number, which can be verified online.
Can the Oracle AI Database 26ai RMAN course be run in-house for our team?
Yes. It can be run in-house for your team, with examples adapted to your own work, such as planning backups and rehearsing recovery for the Oracle releases and database layout your team already runs. You can request a quotation from this page, or contact us on LINE @itgenius or by phone on 02-570-8449.
Much of the work in organisations that run Oracle Database involves multi-step data processing, such as payroll calculations, end-of-day closing or moving data between tables. Written as one SQL statement at a time from the application side, this work is…
Many problems in business systems start with the database structure: duplicated data that is changed in one place but not another, reports that pull figures that do not match, or systems where new features are hard to add because the tables were related…