Linux Swap Management — Swap Files, Partitions & Tuning Guide

In this tutorial, you'll learn about Linux Swap Management. We cover key concepts, practical examples, and best practices to help you understand and apply this topic effectively.

Swap is disk space used as an overflow for RAM — when physical memory fills up, the kernel moves inactive pages to swap, freeing RAM for active processes and preventing out-of-memory crashes.

What You'll Learn

How to create and remove swap files and partitions, tune swappiness and cache pressure, use zswap and zram for compressed in-memory swap, monitor swap usage, and configure swap for different workload types.

Why Swap Matters

A server without swap is one memory spike away from an OOM kill. Swap provides a safety net: the kernel can page out cold memory to disk, keeping the system responsive under pressure. Properly configured swap also enables hibernation (suspend-to-disk). Durga Antivirus Pro's scan servers use zram-backed swap to absorb memory spikes during signature updates without thrashing SSDs.

Learning Path

flowchart LR
  A[Filesystems Compared] --> B[Swap Management
You are here]
  B --> C[Performance Tuning]
  B --> D[System Rescue]
  style B fill:#f90,color:#fff

Swap Files vs Swap Partitions

Modern Linux supports both equally well. Swap files are preferred for flexibility.

Creating a Swap File

# Create a 4GB swap file
sudo dd if=/dev/zero of=/swapfile bs=1M count=4096 status=progress

# Set correct permissions (must be 600)
sudo chmod 600 /swapfile

# Format as swap
sudo mkswap /swapfile

# Enable it
sudo swapon /swapfile

# Make permanent
echo '/swapfile none swap sw 0 0' | sudo tee -a /etc/fstab

Expected output:

$ sudo swapon --show
NAME       TYPE SIZE USED PRIO
/swapfile  file   4G   0B   -2

Creating a Swap Partition

# Create a swap partition (using parted as example)
sudo parted /dev/sdb mklabel gpt
sudo parted /dev/sdb mkpart primary linux-swap 1MiB 100%

# Format as swap
sudo mkswap /dev/sdb1

# Get UUID
sudo blkid /dev/sdb1

# Enable
sudo swapon /dev/sdb1

# Make permanent (using UUID)
echo 'UUID=abc123-... none swap sw 0 0' | sudo tee -a /etc/fstab

Managing Multiple Swap Areas

You can have multiple swap areas with different priorities. The kernel uses higher-priority swap first:

# Add swap with priority
sudo swapon -p 10 /swapfile    # Higher priority (used first)
sudo swapon -p 5 /dev/sdb1     # Lower priority (used when higher fills)

# Check priority and usage
swapon --show

# Remove a swap area
sudo swapoff /swapfile
sudo swapoff /dev/sdb1

Expected output:

$ swapon --show
NAME       TYPE   SIZE  USED PRIO
/swapfile  file     4G  1.2G   10
/dev/sdb1  partition 8G  0.5G    5

Swap Monitoring

# Overall swap usage
free -h

# Per-process swap usage
for file in /proc/*/status; do
    awk '/VmSwap|Name/{printf $2 " " $3}END{ print ""}' $file 2>/dev/null
done | sort -k2 -n -r | head -10

# Using smem
sudo smem -t -s swap

# Check swap activity with vmstat
vmstat 1 10

# With sar
sar -S 1 3

Expected free output:

$ free -h
               total        used        free      shared  buff/cache   available
Mem:            15G         12G        891M        245M        2.0G        2.3G
Swap:           11G        1.7G        9.3G

Swappiness Tuning

Swappiness (0–100) controls how aggressively the kernel swaps. Lower values keep data in RAM longer.

# Check current swappiness
cat /proc/sys/vm/swappiness
# Default: 60

# Set swappiness temporarily
sudo sysctl vm.swappiness=10

# Set permanently
echo 'vm.swappiness=10' | sudo tee /etc/sysctl.d/99-swap.conf

Swappiness Recommendations

VFS Cache Pressure

# Default: 100
cat /proc/sys/vm/vfs_cache_pressure

# Reduce cache reclaim frequency
sudo sysctl vm.vfs_cache_pressure=50

# Make permanent
echo 'vm.vfs_cache_pressure=50' | sudo tee -a /etc/sysctl.d/99-swap.conf

Lower values (50) keep dentry and inode caches longer at the expense of other memory uses. Higher values (200) reclaim cache more aggressively.

zswap — Compressed In-Memory Swap

zswap intercepts pages being swapped out, compresses them, and stores them in a RAM pool. If the pool fills, pages are written to disk swap.

# Enable zswap (at boot)
echo 'zswap.enabled=1 zswap.compressor=zstd zswap.max_pool_percent=20' | \
    sudo tee /etc/default/grub.d/zswap.cfg

# Update grub and reboot
sudo update-grub
sudo reboot

# Check zswap stats
cat /sys/kernel/debug/zswap/

zram — Compressed RAM Block Device

zram creates a compressed block device in RAM that acts as swap entirely in memory — no disk involved. Ideal for systems with limited RAM (VMs, Raspberry Pi).

# Install
sudo apt install zram-tools

# Configure
echo 'ALGO=zstd' | sudo tee -a /etc/default/zramswap
echo 'PERCENT=50' | sudo tee -a /etc/default/zramswap

# Enable
sudo systemctl enable --now zramswap

# Check
zramctl

Expected output:

$ zramctl
NAME       ALGORITHM DISKSIZE  DATA  COMPR  TOTAL STREAMS MOUNTPOINT
/dev/zram0 zstd        7.5G  2.3G  987M  1001M       4 [SWAP]

The compression ratio here is 2.3G of data compressed to ~1001M — nearly 2.3x savings.

Swap and Hibernation

Hibernation (suspend-to-disk) saves RAM contents to swap. The swap area must be large enough to hold all of RAM.

# Check if the system can hibernate
sudo systemctl hibernate

# Configure resume kernel parameter
sudo sed -i 's/GRUB_CMDLINE_LINUX="/GRUB_CMDLINE_LINUX="resume=UUID=SWAP_UUID/' /etc/default/grub
sudo update-grub

# Resume offset for swap files (not partitions)
sudo btrfs inspect-internal map-swapfile -r /swapfile  # If on Btrfs

Common Errors

1. swapon Failed: "Invalid argument"

The swap file has wrong permissions or is on a filesystem that does not support swap (e.g., Btrfs without specific configuration). Fix: sudo chmod 600 /swapfile and ensure the filesystem supports swap files.

2. Out of Memory (OOM) With Swap Available

The OOM killer activates when memory pressure is extreme. Swap may exist but the system is thrashing — spending all CPU swapping pages in and out. Add more RAM or reduce workload.

3. Swap File on Btrfs

Btrfs supports swap files only with specific limitations: no compression, no copy-on-write, and no snapshots on the file. Create the file with chattr +C beforehand to disable CoW.

4. Swappiness=0 Does Not Disable Swap

Swappiness=0 means the kernel avoids swapping unless memory pressure is extreme. It does not disable swap entirely. To truly disable swap, remove or turn off all swap areas.

5. SSD Wear From Excessive Swapping

Frequent swapping wears SSDs. Mitigate with: (1) increase RAM, (2) use zswap/zram to keep swapped pages compressed in RAM, (3) set swappiness to 1 on SSD-only systems.

6. Hibernation Fails Because Swap Is Too Small

The swap area must be at least as large as the total installed RAM. Use a dedicated swap partition for hibernation instead of a swap file.

7. High Swap Usage Despite Available RAM

This usually indicates misconfigured swappiness or an application leaking memory. Use smem to find the culprit Process. Check vmstat 1 for si (swap in) and so (swap out) columns.

Practice Questions

1. What is the difference between swap and zram? Swap writes pages to disk. zram compresses pages in RAM and uses them as swap — no disk I/O. zram is faster but uses RAM for compressed storage.

2. What swappiness value is recommended for database servers? 1–10. Databases benefit from keeping their working set in RAM and avoiding swap latency.

3. How do you check which processes are using swap? Read /proc/*/status and grep for VmSwap, or use sudo smem -t -s swap.

4. What is the advantage of a swap file over a swap partition? Swap files can be created, resized, and removed without repartitioning. They work on most modern filesystems and support snapshots of the containing volume.

5. How does zswap differ from zram? zswap intercepts swapped pages, compresses them, and stores them in a RAM cache. If the cache fills, pages spill to disk swap. zram creates a compressed RAM block device used directly as swap, never touching disk.

Challenge: Set up a system with 2GB of zram swap (priority 100), an 8GB swap file on SSD (priority 50), and a 4GB swap partition on HDD (priority 10). Generate memory pressure with stress --vm 4 --vm-bytes 10G --timeout 60. Monitor swap-in/swap-out rates with vmstat 1, compression ratios with zramctl, and verify the priority order. Then tune swappiness to minimize SSD writes.

What's Next

Built by the developers of Doda Browser, DodaZIP, and Durga Antivirus Pro. Updated 2026-06-24.