A Redo Log is a file that records every change made to the Oracle database , whether the transaction is committed or not. Whenever a user performs an operation like: - INSERT INSERT - UPDATE UPDATE - DELETE DELETE - DDL statements DDL statements Oracle records the change in the Redo Log Buffer , and then LGWR (Log Writer process) writes it to the redo log files.

- Stores change vectors (not actual data) Stores change vectors (not actual data) - Written sequentially Written sequentially - Required for instance recovery Required for instance recovery - Works in a circular fashion Works in a circular fashion - Overwritten automatically Overwritten automatically - Mandatory for all Oracle databases Mandatory for all Oracle databases

- Redo Log Buffer (SGA) Redo Log Buffer (SGA) - Online Redo Log Files Online Redo Log Files - Log Groups and Members Log Groups and Members - LGWR Process LGWR Process

- Crash recovery Crash recovery - Instance recovery Instance recovery - Protects against memory failures Protects against memory failures - Ensures database consistency Ensures database consistency Without redo logs, Oracle cannot recover from crashes .

An Archive Log is a copy of a filled redo log file that is saved to a separate location before the redo log is reused. Archive logs are generated only when the database is running in ARCHIVELOG mode . The ARCH (Archiver) process is responsible for copying filled redo logs to archive destinations.

- Stores historical redo data Stores historical redo data - Created after redo log switch Created after redo log switch - Used for media recovery Used for media recovery - Not overwritten automatically Not overwritten automatically - Required for point-in-time recovery Required for point-in-time recovery - Mandatory for Data Guard & Standby databases Mandatory for Data Guard & Standby databases

- ARCH Process ARCH Process - Archive Destinations Archive Destinations - Archive Log Files Archive Log Files - FRA (Fast Recovery Area) FRA (Fast Recovery Area)

- User performs a transaction User performs a transaction - Changes go to Redo Log Buffer Changes go to Redo Log Buffer - LGWR writes changes to Redo Log File LGWR writes changes to Redo Log File - Redo Log becomes full Redo Log becomes full - Log switch occurs Log switch occurs - ARCH copies redo log to archive location ARCH copies redo log to archive location - Redo log is reused Redo log is reused This mechanism ensures zero data loss recovery capability when properly configured.

- Oracle uses Redo Logs to recover uncommitted transactions Oracle uses Redo Logs to recover uncommitted transactions - Archive logs are not involved Archive logs are not involved

- Datafiles are restored from backup Datafiles are restored from backup - Archive logs are applied Archive logs are applied - Database recovers to the latest point Database recovers to the latest point

- Archive logs are shipped to standby Archive logs are shipped to standby - Applied for synchronization Applied for synchronization

- RMAN uses archive logs RMAN uses archive logs - Restores database to a specific SCN or time Restores database to a specific SCN or time

❌ No point-in-time recovery ❌ No Data Guard ❌ Risk of data loss ❌ Limited backup strategy ✔ Only suitable for test or non-critical environments

- Always enable ARCHIVELOG mode in production Always enable ARCHIVELOG mode in production - Monitor archive log space (FRA) Monitor archive log space (FRA) - Implement regular archive log backups Implement regular archive log backups - Delete archive logs after backup using RMAN Delete archive logs after backup using RMAN - Multiplex redo logs for high availability Multiplex redo logs for high availability

Simple way to explain in interviews: Redo Logs capture live database changes, while Archive Logs store those changes permanently for recovery.