How to Set Up and Manage File-Based Swap on Linux

When you need additional swap space but all partitions are allocated, creating a swap file lets you add capacity without repartitioning. This is especially useful on systems with fixed disk layouts or cloud instances where you can’t modify the underlying disk structure.

Creating the Swap File

Start by allocating space. fallocate is the fastest option on most modern filesystems:

fallocate -l 2G /swapfile

If you’re on btrfs, NFS, or another filesystem where fallocate fails, use dd instead:

dd if=/dev/zero of=/swapfile bs=1M count=2048

The bs=1M block size is significantly faster than smaller sizes for large files. Verify creation:

ls -lh /swapfile

Set Proper Permissions

Swap files can contain sensitive data in memory dumps, so restrict access:

chmod 600 /swapfile
chown root:root /swapfile

The 600 permission (read/write for root only) is the standard. Don’t use 700—it’s unnecessarily permissive.

Initialize the Swap Space

Format the file as swap:

mkswap /swapfile

You’ll see output like:

Setting up swapspace version 1, size = 2 GiB (2147479552 bytes)
no label, UUID=f30730f3-eb1d-4d30-a088-a0274f3c2f3e

Ignore warnings about bootbits—these are harmless.

Enable the Swap File

Activate it immediately:

swapon /swapfile

Verify it’s active:

swapon --show

Output shows:

NAME      TYPE SIZE USED PRIO
/swapfile file 2.0G   0B  -2

Check overall memory and swap status:

free -h

Make Swap Persistent Across Reboots

Add an entry to /etc/fstab:

/swapfile none swap sw 0 0

Use none as the mount point, not swap—this is the modern convention. Test the configuration without rebooting:

swapon --all --verbose

This validates that fstab parses correctly.

Managing Multiple Swap Files

To add additional swap alongside existing swap:

fallocate -l 4G /swapfile2
chmod 600 /swapfile2
mkswap /swapfile2
swapon /swapfile2

Then append to /etc/fstab:

echo "/swapfile2 none swap sw 0 0" >> /etc/fstab

Each swap file can have different priorities. Lower priority numbers are used first. To set a priority when enabling swap:

swapon -p 10 /swapfile2

Then update /etc/fstab with the priority option:

/swapfile2 none swap sw,pri=10 0 0

Removing Swap

To disable a swap file:

swapoff /swapfile

Remove its line from /etc/fstab, then delete the file:

rm /swapfile

Monitoring Swap Usage

Check overall swap consumption:

grep Swap /proc/meminfo

For per-process swap usage, use:

for pid in $(pgrep -P 1); do 
  label=$(cat /proc/$pid/comm 2>/dev/null)
  swap=$(grep VmSwap /proc/$pid/status 2>/dev/null | awk '{print $2}')
  [ -n "$swap" ] && [ "$swap" != "0" ] && echo "$label (PID $pid): ${swap}kB"
done

Or use smem if installed:

smem -s swap -r

Watch swap activity in real-time with iostat (from sysstat package):

iostat -x 1 | grep -E "Device|sda"

Performance Considerations

Filesystem choice matters. File-based swap performs nearly identically to partition-based swap on ext4 and XFS. Modern NVMe disks make swap files practical even for performance-critical systems.

Btrfs requires care. Copy-on-Write (CoW) conflicts with swap. Disable CoW on the swap file:

chattr +C /swapfile

Do this before running mkswap.

Avoid network filesystems. NFS and other network-mounted filesystems are too slow for swap. Always place swap on local storage.

Tune swappiness. The vm.swappiness setting (0–100) controls how aggressively the kernel uses swap. Check the current value:

cat /proc/sys/vm/swappiness

Set it temporarily:

sysctl vm.swappiness=30

For servers, 30 is a reasonable default. For desktops, higher values (50–60) may feel more responsive. Make it permanent in /etc/sysctl.conf:

vm.swappiness=30

Size recommendations. For modern systems with adequate RAM, swap equal to 25% of physical RAM is usually sufficient. For servers with 64GB+ RAM, even 4–8GB of swap handles edge cases. Virtual machines and containers benefit from slightly more swap due to unpredictable load spikes.