Why Journaling Keeps File Systems Safe: Write-Ahead Logging Explained

File systems must prevent data corruption caused by power loss or crashes, which occurs because file writes are not atomic and involve both user data and metadata such as superblock, inode bitmap, inode, and data block bitmap.

A simplified write operation includes three steps: allocating a data block from the data block bitmap, adding a pointer to that block in the inode, and writing user data into the block.

Allocate a data block from the data block bitmap.

Add a pointer to the data block in the inode.

Write user data into the data block.

If any step is interrupted, various inconsistencies arise: data corruption, metadata inconsistency, wasted space, or orphaned data.

Journaling file systems solve this by recording the intended steps (a transaction) in a separate log before performing the actual writes. After the log is safely stored, the system proceeds to write metadata and user data (checkpoint). If a crash occurs, the log is replayed on the next mount, restoring consistency.

To handle power loss during log writing, each log entry is written with an end‑marker; only entries with a valid end‑marker are considered complete, ensuring atomicity despite block‑size limitations.

The journal lifecycle includes four phases:

Journal write: record the transaction in the log.

Journal commit: write the end‑marker after the log entry is fully stored.

Checkpoint: perform the actual writes of metadata and user data to the file system.

Free: reclaim the space occupied by the used log entries.

Data journaling records both metadata and user data in the log, but this doubles write traffic and reduces efficiency. Metadata (or ordered) journaling records only metadata in the log, writing user data directly before the journal entry, which improves performance while still guaranteeing consistency.

Linux’s EXT3 file system supports both data journaling and ordered (metadata) journaling, allowing administrators to choose the appropriate mode.

Reference: Crash Consistency – FSCK and Journaling.