WAL Replay

Definitions

1. The crash recovery process where a database reads its write-ahead log from the last checkpoint and re-applies committed transactions that weren't yet written to data files. Ensures no committed data is lost even after an unexpected shutdown. The reason databases can guarantee durability despite crashes.

   In plain English: The process of replaying the change log after a database crash to restore any committed data that wasn't fully saved yet.

   Example: "PostgreSQL crashed mid-checkpoint but recovered cleanly — WAL replay re-applied 3 minutes of committed transactions on restart."

Related Terms

• WAL
• Consistency
• Ecto
• JDBC
• NoSQL
• Query
• Write-Ahead Log
• Transaction Log