Write Ahead Log(WAL)

Definition:

Write Ahead Log (WAL) is a technique used in distributed systems where all changes are first written to a durable log before being applied to the actual data. This ensures data consistency, enables crash recovery, and helps maintain data integrity across multiple nodes. The key principle is: “log first, then apply changes.”

Example:

• PostgreSQL Database Cluster

-- User executes: UPDATE users SET balance = 500 WHERE id = 123;

-- Step 1: WAL entry written to disk FIRST
WAL Record:
{
  "lsn": "0/1B000028",
  "type": "UPDATE",
  "table": "users",
  "old_data": {"id": 123, "balance": 400},
  "new_data": {"id": 123, "balance": 500},
  "transaction_id": "TXN_456"
}

-- Step 2: Only AFTER WAL is on disk, apply to database
-- Step 3: Eventually, WAL gets replicated to standby servers

Recovery Process:

Primary Server Crashes!

Standby Server:
1. Reads WAL from shared storage
2. Replays all WAL entries since last checkpoint
3. Becomes new primary
4. No data lost! ✓

WAL in Distributed Systems:

┌─────────────┐    ┌─────────────┐    ┌─────────────┐
│   Client    │    │   Node 1    │    │   Node 2    │
└─────────────┘    └─────────────┘    └─────────────┘
       │                   │                   │
       │ 1. Write Request  │                   │
       ├──────────────────>│                   │
       │                   │ 2. Write to WAL   │
       │                   ├──────────────────>│
       │                   │ 3. WAL Synced     │
       │                   │<──────────────────┤
       │                   │ 4. Apply Changes  │
       │                   │ 5. Apply Changes  │
       │                   ├──────────────────>│
       │ 6. Success        │                   │
       │<──────────────────┤                   │