Database Corruption Recovery Advisor

Database corruption is one of the most feared incidents in data management — it can range from a single corrupt page affecting a small table to systemic storage failures threatening an entire instance. This AI role provides expert guidance for identifying the scope of corruption, assessing recovery options, and executing the safest possible path to data restoration.

The Database Corruption Recovery Advisor helps you understand corruption error messages from PostgreSQL (invalid page header, checksum mismatch), SQL Server (page torn, 823/824/825 errors), MySQL InnoDB (tablespace corruption messages), and Oracle (ORA-01578 block corruption errors). It explains what each error type means in terms of which data structures are affected and what the likely cause is — whether hardware failure, storage subsystem issues, sudden power loss, or software bugs.

For each corruption scenario, the role walks you through a structured assessment: determine whether the corruption is in a critical system object or a user table, check whether backups are clean and current, evaluate whether native repair tools (DBCC CHECKDB repair options, InnoDB force recovery modes, pg_resetwal) are appropriate, and determine the full extent of data loss if restoration from backup is required.

This role is particularly valuable because corruption recovery is high-stakes and non-reversible actions — like enabling InnoDB force recovery or running DBCC CHECKDB with repair options — can cause additional data loss if used incorrectly. The advisor ensures you understand the implications of every step before taking it and always presents the most conservative option first.

Ideal users include DBAs facing unexpected corruption alerts, engineers managing databases on aging hardware, and teams recovering from infrastructure incidents that affected storage. This role turns a terrifying situation into a methodical, documented recovery process.