e2fsprogs 1.47.2

ext4 is still the default filesystem on most Linux installations, and e2fsprogs is the toolbox for it.

The tools

e2fsck checks and repairs. fsck dispatches to it for ext filesystems.

mke2fs creates them, usually invoked as mkfs.ext4.

resize2fs grows or shrinks, which is what lvextend -r calls for you.

tune2fs changes parameters on an existing filesystem without recreating it.

dumpe2fs shows the superblock and block group information.

Checking and repairing

sudo umount /dev/sdb1
sudo fsck -n /dev/sdb1     # report only, change nothing
sudo fsck -y /dev/sdb1     # answer yes to all repairs

Never on a mounted filesystem. The kernel is modifying it, so fsck reads a moving target and any repair can conflict with in-flight changes. Running it on a mounted filesystem can corrupt a healthy one.

Start with -n. It tells you the scale of the problem before you commit to a destructive repair. Our fsck guide covers the whole procedure including imaging first.

tune2fs is the underused one

sudo tune2fs -l /dev/sdb1              # show everything
sudo tune2fs -L mydata /dev/sdb1       # set a label
sudo tune2fs -c 30 /dev/sdb1           # check every 30 mounts
sudo tune2fs -m 1 /dev/sdb1            # reserved blocks down to 1%

That last one matters on a large data volume. ext4 reserves 5% of the filesystem for root by default, which prevents a full disk from making the system unbootable. On a 8TB data drive that is 400GB reserved for no reason, and dropping it to 1% reclaims most of it.

Leave the default on a root filesystem.

Resizing

sudo resize2fs /dev/mapper/vg0-data              # grow to fill the device
sudo resize2fs /dev/mapper/vg0-data 100G         # explicit size

ext4 can grow online while mounted. Shrinking requires unmounting and shrinking the filesystem before the volume, and getting that order wrong destroys data. Our LVM guide covers the sequence.

Creating with sensible options

sudo mkfs.ext4 -L mydata -m 1 /dev/sdb1
sudo mkfs.ext4 -T largefile4 /dev/sdb1     # fewer inodes, for large files

Inode count is fixed at creation. A filesystem holding millions of tiny files can exhaust inodes while showing free space, and the only fix is recreating it.

Verify Your Download

Source releases on kernel.org are published with GPG signatures.