How to corrupt an SQLite database file

6 days ago

SQLite databases are highly resistant to corruption, with automatic rollback on crashes, but they are not immune.
Corruption can occur due to file descriptor misuse, such as overwriting database files with unrelated data.
Background backup processes copying during transactions can create corrupt backups; safe methods include sqlite3_rsync, VACUUM INTO, and the backup API.
Moving, deleting, or renaming hot journal files (like *-journal or *-wal) after a crash prevents recovery and may corrupt the database.
File locking issues, especially on network filesystems or with POSIX advisory locking quirks, can lead to corruption if multiple processes write concurrently.
Using multiple links (hard or symbolic) to a database file can cause undefined behavior and potential corruption.
Forking a process with an open SQLite connection and using it in the child can result in locking problems and corruption.
Storage devices lying about sync operations (e.g., USB sticks) may cause out-of-order writes, leading to corruption, mitigated by using WAL mode.
Disabling SQLite protections like PRAGMA synchronous=OFF increases corruption risk during crashes or power failures.
Bugs in SQLite, though rare, have historically caused corruption, such as the WAL-reset bug and issues with expression indexes.

Hasty Briefsbeta