2.2Managing Filesystems

A filesystem is not "create and forget"—it must be checked, tuned, and backed up throughout its operational life. ext4, XFS, and Btrfs each embody a different design philosophy with its own tool family, so mapping "which command belongs to which filesystem" accurately matters both in practice and on the exam.

2.2.1Creation and integrity checks

• mkfs is the generic command to create a filesystem; the real work happens in per-type implementations like mkfs.* (mkfs.ext4, mkfs.xfs, mkfs.btrfs), with mkfs -t ext4 acting as a wrapper that calls them. Creation is a destructive operation that discards any existing data.
• fsck is the generic integrity-check-and-repair command, implemented per type as fsck.* (e.g., fsck.ext4). Running it against a mounted filesystem is generally unsafe (risk of corruption); the standard practice for the root filesystem is to check it in single-user or rescue mode at boot.

2.2.2ext-family tuning/backup, XFS, Btrfs, and SMART

• tune2fs adjusts ext2/3/4 parameters (label, reserved-block percentage, check interval, etc.) while the filesystem stays mounted. dumpe2fs displays detailed information such as the superblock and block-group data (read-only, no changes).
• dump/restore form the traditional pair for full, differential, and incremental backup/restore of ext-family filesystems (dump levels manage generations).
• XFS-specific tools (dot-separated naming): xfs_info (show layout), xfs_check (legacy check), xfs_repair (repairs while unmounted; fsck.xfs is effectively a no-op), and xfsdump/xfsrestore (the XFS equivalents of dump/restore).
• Btrfs splits one filesystem into independent namespaces via subvolumes, and its snapshots use copy-on-write (CoW) for instant, space-efficient point-in-time captures. Operations are unified under btrfs subcommands like btrfs subvolume create/list and btrfs subvolume snapshot; migrating an existing ext filesystem in place uses btrfs-convert.
• smartd is a background daemon that continuously monitors SMART (Self-Monitoring, Analysis and Reporting Technology)-capable disks (logging/notifying on threshold breaches). smartctl is the tool for manual queries and self-tests (smartctl -a /dev/sda lists attributes; smartctl -t short starts a test).

Mapping the tradeoffs of the three filesystems makes exam context easier to read. ext4 offers proven maturity and broad track record, backed by a settled tool family (tune2fs, dumpe2fs, dump/restore)—a safe default for straightforward use cases. XFS is designed for large capacity and highly parallel I/O; xfs_repair supports largely online operations day-to-day (though the repair itself requires unmounting), and xfsdump/xfsrestore provide level-based backups—popular for large file servers and databases. Btrfs differentiates itself with CoW-based subvolumes and snapshots, making it easy to "snapshot before a production change, then instantly roll back if something goes wrong." To migrate an existing ext4 filesystem to Btrfs without rebuilding data, use btrfs-convert—but note the converted filesystem retains rollback information pointing back to the original ext4 image, which must be explicitly discarded once you commit to the conversion. Disk-level physical health is monitored on a separate axis via SMART: run smartd continuously to catch threshold breaches, and when suspicious signs appear (e.g., rising reallocated sector counts), run a detailed self-test with smartctl -t long before proceeding to disk replacement if warranted—the standard workflow for managing hardware-failure precursors.

2.2.3Section summary

• mkfs/fsck are generic, with per-type .type implementations. fsck requires unmounting (root: single-user mode); tune2fs adjusts while mounted, dumpe2fs is read-only
• XFS = xfs_repair (unmount required), xfsdump/xfsrestore / Btrfs = CoW subvolumes, snapshots, btrfs-convert / SMART = smartd daemon, smartctl manual query