Filesystems (XFS/EXT4)

XFS and EXT4 are the two most commonly used filesystems in Enterprise Linux. The RHEL/CentOS family uses XFS by default, while EXT4 is more common in the Debian/Ubuntu family. Understanding their differences and respective management tools is essential for system administrators.

XFS vs. EXT4 Comparison

Feature	XFS	EXT4
Default in	RHEL 7+ / CentOS 7+	Debian / Ubuntu
Maximum filesystem capacity	8 EB	1 EB
Maximum single file size	8 EB	16 TB
Online expansion	Supported (grow only)	Supported (grow only)
Online shrink	Not supported	Not supported online, but supported offline
Journal type	Metadata journal	Metadata + data journal
Defragmentation	xfs_fsr	e4defrag
Repair tool	xfs_repair	e2fsck
Best suited for	Large files, high-concurrency I/O	General purpose, many small files
File deletion performance	Excellent	Good

Tip

If you are using the RHEL/CentOS/Rocky Linux/AlmaLinux family, it is recommended to use the default XFS. If you need to shrink a filesystem, EXT4 is the more suitable choice.

Creating Filesystems

Creating an XFS Filesystem

Basic creation:

sudo mkfs.xfs /dev/sdb1

Specifying a label and block size:

sudo mkfs.xfs -L appdata -b size=4096 /dev/sdb1

Force re-format (overwrite an existing filesystem):

sudo mkfs.xfs -f /dev/sdb1

Common mkfs.xfs parameters:

Parameter	Description
-f	Force format
-L label	Set the filesystem label
-b size=N	Set the block size (default 4096)
-i size=N	Set the inode size
-d agcount=N	Set the number of allocation groups

Creating an EXT4 Filesystem

Basic creation:

sudo mkfs.ext4 /dev/sdb1

Specifying a label, block size, and inode ratio:

sudo mkfs.ext4 -L appdata -b 4096 -i 8192 /dev/sdb1

Common mkfs.ext4 parameters:

Parameter	Description
-L label	Set the filesystem label
-b blocksize	Set the block size
-i bytes-per-inode	Allocate one inode per this many bytes
-N number	Specify the total number of inodes
-m percentage	Percentage of space reserved for root (default 5%)
-O feature	Enable a specified feature

Note

EXT4 reserves 5% of space for the root user by default. For data-only partitions, you can use -m 0 or -m 1 to reduce the reserved space and gain more usable capacity.

XFS Management Tools

xfs_info - Viewing XFS Filesystem Information

sudo xfs_info /dev/sdb1

Or view via mount point:

sudo xfs_info /app

Example output:

meta-data=/dev/sdb1              isize=512    agcount=4, agsize=3276800 blks
         =                       sectsz=512   attr=2, projid32bit=1
         =                       crc=1        finobt=1, sparse=1, rmapbt=0
         =                       reflink=1    bigtime=1 inobtcount=1
data     =                       bsize=4096   blocks=13107200, imaxpct=25
         =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0, ftype=1
log      =internal log           bsize=4096   blocks=6400, version=2
         =                       sectsz=512   sunit=0 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0

Key fields explained:

• bsize: Block size
• blocks: Total number of blocks
• agcount: Number of allocation groups
• isize: Inode size

xfs_repair - Repairing an XFS Filesystem

Caution

xfs_repair can only be run when the filesystem is unmounted. Unmount the filesystem before repairing.

Basic repair procedure:

1. Unmount the filesystem:

   sudo umount /app

2. First check whether the log can be replayed:

   sudo xfs_repair -n /dev/sdb1

   The -n flag runs in check-only mode without making repairs.

3. Perform the repair:

   sudo xfs_repair /dev/sdb1

4. If a corrupted log prevents repair, clear the log and retry:

   sudo xfs_repair -L /dev/sdb1

   -L discards uncommitted transactions in the log, which may result in minor data loss. Use only when necessary.

5. Remount:

   sudo mount /dev/sdb1 /app

xfs_growfs - Online XFS Expansion

After a logical volume or partition has been enlarged, expand the XFS filesystem to use the new space:

# Expand to fill all available space on the device at the mount point
sudo xfs_growfs /app

You can also specify a new block count:

sudo xfs_growfs -D 26214400 /app

Danger

XFS filesystems do not support shrinking. If you need to reduce storage space, you must back up the data, delete the filesystem, create a smaller one, and then restore the data.

Other XFS Tools

# Defragmentation
sudo xfs_fsr /app

# Back up an XFS filesystem
sudo xfsdump -l 0 -f /backup/app.dump /app

# Restore an XFS filesystem
sudo xfsrestore -f /backup/app.dump /app

# Freeze a filesystem (for consistent snapshots)
sudo xfs_freeze -f /app
# Unfreeze
sudo xfs_freeze -u /app

EXT4 Management Tools

tune2fs - Viewing and Adjusting EXT4 Parameters

View detailed filesystem information:

sudo tune2fs -l /dev/sdb1

Common adjustment operations:

# Change the filesystem label
sudo tune2fs -L newlabel /dev/sdb1

# Adjust the reserved space percentage (from default 5% to 1%)
sudo tune2fs -m 1 /dev/sdb1

# Set maximum mount count before a forced check (0 to disable)
sudo tune2fs -c 0 /dev/sdb1

# Set the check interval (0 to disable time-based checks)
sudo tune2fs -i 0 /dev/sdb1

# Enable a filesystem feature
sudo tune2fs -O has_journal /dev/sdb1

# View the UUID
sudo tune2fs -l /dev/sdb1 | grep UUID

e2fsck - Checking and Repairing EXT4

Caution

e2fsck can only be run on an unmounted filesystem. Running it on a mounted filesystem may cause severe data corruption.

1. Unmount the filesystem:

   sudo umount /data

2. Check the filesystem (without repairing):

   sudo e2fsck -n /dev/sdb1

3. Automatically repair all issues:

   sudo e2fsck -y /dev/sdb1

4. Force check (even if the filesystem is marked clean):

   sudo e2fsck -f /dev/sdb1

5. Remount:

   sudo mount /dev/sdb1 /data

Common e2fsck parameters:

Parameter	Description
-n	Check only, do not repair
-y	Automatically answer yes to all questions
-f	Force check
-p	Automatically repair safe issues
-v	Verbose output

resize2fs - Resizing an EXT4 Filesystem

Expanding an EXT4 Filesystem

Online expansion (while the filesystem is mounted):

# First extend the logical volume or partition
sudo lvextend -L +20G /dev/vg_data/lv_data

# Then expand the filesystem to fill the entire device
sudo resize2fs /dev/vg_data/lv_data

Expand to a specific size:

sudo resize2fs /dev/vg_data/lv_data 80G

Shrinking an EXT4 Filesystem

Danger

Shrinking a filesystem carries a risk of data loss! Be sure to back up your data first. The shrink operation must be performed with the filesystem unmounted.

1. Unmount the filesystem:

   sudo umount /data

2. Force check the filesystem:

   sudo e2fsck -f /dev/vg_data/lv_data

3. Shrink the filesystem:

   sudo resize2fs /dev/vg_data/lv_data 30G

4. Shrink the logical volume (must match or be slightly larger than the filesystem size):

   sudo lvreduce -L 30G /dev/vg_data/lv_data

5. Remount and verify:

   sudo mount /dev/vg_data/lv_data /data
   df -hT /data

Practical Tips

Checking the Filesystem Type

# Method 1: lsblk
lsblk -f

# Method 2: blkid
sudo blkid /dev/sdb1

# Method 3: df (for mounted filesystems)
df -T /app

# Method 4: file command
sudo file -sL /dev/sdb1

Checking Filesystem Usage

# View usage of all mount points
df -hT

# View inode usage
df -i

# View disk usage for a specific directory
du -sh /app/*

Changing a Filesystem Label

# XFS
sudo xfs_admin -L newlabel /dev/sdb1

# EXT4
sudo tune2fs -L newlabel /dev/sdb1

Changing a Filesystem UUID

# XFS (must be unmounted)
sudo xfs_admin -U generate /dev/sdb1

# EXT4
sudo tune2fs -U random /dev/sdb1

Hands-On: Creating and Managing Dual Filesystems

The following example creates XFS and EXT4 filesystems on two partitions and performs basic management operations.

1. Create two partitions (assuming /dev/sdb is already partitioned):

   lsblk /dev/sdb
   # sdb1 - 50G (for XFS)
   # sdb2 - 50G (for EXT4)

2. Create the filesystems:

   sudo mkfs.xfs -L xfs_data /dev/sdb1
   sudo mkfs.ext4 -L ext4_data -m 1 /dev/sdb2

3. Mount and verify:

   sudo mkdir -p /data/xfs /data/ext4
   sudo mount /dev/sdb1 /data/xfs
   sudo mount /dev/sdb2 /data/ext4
   df -hT /data/xfs /data/ext4

4. View filesystem information:

   # XFS information
   sudo xfs_info /data/xfs
   
   # EXT4 information
   sudo tune2fs -l /dev/sdb2 | head -30

5. Configure auto-mounting:

   sudo cp /etc/fstab /etc/fstab.bak
   cat <<EOF | sudo tee -a /etc/fstab
   LABEL=xfs_data   /data/xfs   xfs   defaults,noatime  0  0
   LABEL=ext4_data  /data/ext4  ext4  defaults,noatime  0  0
   EOF
   sudo mount -a

Mastering XFS and EXT4 management tools is fundamental to Enterprise Linux operations. Choosing the right filesystem for your workload and making proper use of inspection and repair tools will effectively safeguard data integrity.