hash-index-survives-restart

Status: IN

Hash-index storage correctly reconstructs its in-memory index from on-disk segments after restart by scanning in ascending order so newer writes overwrite older entries, with fsync-controlled durability for each write.

Justifications

Restart recovery is correct for normal operation but a crash during compaction leaves segment state that rebuild cannot handle consistently

Depends on (SL): bitcask-rebuild-sorts-ascending, bitcask-recovery-order-is-ascending, hash-index-sync-writes-controls-fsync

Unless: bitcask-compact-not-crash-safe

JSON