The four phases of processing

The recover1 program’s processing consists of up to four separate phases, which are run in the following order:

1. Integrity Scan. If the Q option or Y option is specified, the Integrity Scan phase is disregarded unless it is forced to occur by the specification of the I option or L option. This phase reads the entire file in a forward direction checking for simple errors, and produces a summary report showing the state of the file and an estimate of the number of records recover1 can recover. The indexed file is not modified during this phase.
2. Repair Blocks. The Repair Blocks phase, which is always run, reads and writes the file in a backward direction repairing corrupt data blocks, converting non-data blocks to empty blocks, and rebuilding some internal file structures.
3. Move Data Blocks. The Move Data Blocks phase is run only when the truncate file

   option (T) is specified. This phase reads and writes parts of the file moving highnumbered data blocks (near the end of the file) to lower-numbered available blocks to maximize the amount of space at the end of the file that can be truncated and returned to the operating system when recover1 finishes.

4. Build Node Blocks. The Build Node Blocks phase, which is always run, reads data blocks and writes node blocks in the file in a forward direction, rebuilding the entire node structure for each key of the file.