PAR2

PAR2, short for Parchive version 2, is a data integrity and error-correction format used to add redundancy to a set of files so that damaged or missing data can be repaired without re-downloading. It is an evolution of the original PAR format and is widely used in distributing large archives, particularly in Usenet and BitTorrent communities where transfers can be error-prone.

The technology works by splitting the target data into blocks and generating additional parity blocks through a Reed-Solomon code. These parity blocks are stored in one or more PAR2 files accompanying the original data. A verifier can check checksums to detect errors, and a repair process uses the parity data to reconstruct missing or corrupted blocks. The amount of recoverable data depends on the number of recovery blocks specified when the PAR2 set is created.

Users typically employ a PAR2 creation tool to generate a recovery set, and a verification or repair tool to test and restore the data. Par2cmdline is a common example of such tooling. PAR2 supports multiple files, variable block sizes, and selectable recovery capacities, enabling targeted repair and efficient verification for large archives.

Limitations include the need for sufficient parity data to cover the expected loss; if damage exceeds the recovery capacity, re-downloading may still be necessary. PAR2 is platform-agnostic and designed to be practical for correcting errors in distributed archives.

See also Parchive, Reed-Solomon error correction, data integrity.