Data Disaster Recovery supports logical backup and physical backup. Use this page to understand how each method works and choose the one that fits your database size and performance requirements.
Backup methods
Logical backup
Logical backup exports database objects — tables, indexes, and stored procedures — by running SQL statements against the database. Common tools include mysqldump (MySQL) and exp/imp (Oracle).
Characteristics:
Backs up at the database or table level, so you can restore individual tables without restoring the entire instance — unlike physical backup, which only restores the full instance.
Does not lock databases during backup, which minimizes impact on database performance.
Backup and restoration are generally slower than physical backup, making it better suited for databases up to 1 TB.
Incremental backup: Data Disaster Recovery reads logs stored in database memory and performs incremental backup in real time. The read speed adjusts automatically based on the log generation rate, preventing I/O performance drops.
Supported by Data Disaster Recovery: Yes. Recommended for databases between 1 MB and 1 TB.
Physical backup
Physical backup copies the actual database files from the operating system. A Data Disaster Recovery backup gateway, installed on the database server, handles the file-level copy. Common tools include XtraBackup (MySQL) and RMAN (Oracle).
Characteristics:
Backs up the entire database instance as a whole — table-level or database-level granularity is not available, unlike logical backup.
Reads data directly from disk, which affects database I/O performance during backup.
Faster than logical backup for both backup and restoration, making it the preferred choice for large databases.
After restoring a physical backup, you can create sandbox instances to test or validate the backup data without affecting production. For details, see Data Disaster Recovery sandbox overview.
Supported by Data Disaster Recovery: Yes. Recommended for databases larger than 1 TB. Requires installing a backup gateway on the database server. For setup instructions, see Install a backup gateway.
Snapshot
Snapshot backup creates a fully usable copy of a dataset using snapshot technology. Snapshots can be stored locally or replicated across servers. Common tools include VERITAS File System, Linux LVM, and NetApp NAS.
Supported by Data Disaster Recovery: No. Data Disaster Recovery does not support snapshot-based backup.
Comparison: logical backup vs. physical backup
| Attribute | Logical backup | Physical backup |
|---|---|---|
| Backup scope | Database or table level | Entire database instance only |
| Full backup mechanism | Splits table data and reads it in parallel using SQL statements | Copies database files from disk using a backup gateway installed on the server |
Incremental backup | Supported — reads database logs from memory in real time; adjusts read speed based on log generation rate | — |
| Effect on I/O | Moderate — reads from disk during full backup | Moderate — reads from disk, affects I/O performance |
| Database locking | No lock during backup | — |
| Backup and restore speed | Slower | Faster in most cases |
| Recommended database size | 1 MB–1 TB | Larger than 1 TB |
| Additional capabilities | — | Supports sandbox instances for testing restored data |
What's next
To check which backup methods are available for your database engine, see Database engines and features.
To learn where backup data is stored, see Built-in storage and OSS.
To set up your first backup, see Create a backup schedule.