ext4 — The Traditional Default

fs1-1 previewed journaling without explaining the mechanism. This chapter delivers that in full, plus everything else that makes ext4 the long-standing default Linux filesystem — and an honest answer to when it's still the right choice.

Journaling — Delivering on Chapter 1's Own Preview

Before making a change to the actual filesystem structure, ext4 first writes a small record of the intended change to a dedicated journal area. If the system crashes mid-operation, on the next mount ext4 replays the journal to either complete or cleanly roll back the incomplete change — avoiding the pre-journaling era's own slow, full-filesystem scan after every unclean shutdown.

• journal — full data and metadata journaling; the safest mode, the slowest
• ordered (the default) — metadata is journaled; actual data is written before its metadata commits — a genuinely good balance
• writeback — metadata only; the fastest, but the weakest guarantee — data itself could still be corrupted after a crash even though the filesystem's own structure stays consistent

Extents — A Real Improvement Over Block Mapping

Older filesystems (ext2/ext3) tracked a file's data location as a long list of individual block pointers — genuinely inefficient for a large, mostly contiguous file, which might need thousands of individual pointers. ext4's own extents represent a large contiguous run of blocks as one compact record — "blocks N through N+1000 belong to this file" — dramatically more efficient for large files, both in metadata size and in reducing fragmentation-related overhead.

Creating and Tuning an ext4 Filesystem

tune2fs is the real, primary tool for adjusting ext4 parameters after creation — labels, reserved-block percentage, forced-check intervals — without needing to reformat and lose everything.

fsck — Checking and Repairing

Checks, and can repair, filesystem consistency. Genuinely important: the check assumes nothing else is actively modifying the filesystem while it runs. For a root filesystem that can't simply be unmounted while running, grub1-9's own chroot-rescue material is directly relevant — boot a live environment, mount the target filesystem there instead, and run fsck from outside it.

When ext4 Is Still the Right Choice

ext4 is mature, extremely well-tested, has the broadest compatibility and tooling support of any Linux filesystem, and has genuinely lower resource overhead than fs1-3's Btrfs or fs1-4's ZFS. The right default when you don't specifically need snapshots, checksumming, or integrated volume management, and simplicity, stability, and raw performance matter most. Btrfs and ZFS aren't free upgrades over ext4 — each trades some of this simplicity for genuinely new capabilities, a real tradeoff this course takes seriously rather than treating ext4 as simply "the old one."

Hands-On Exercises