Swap Memory

Swap space is an overflow buffer on your physical storage (HDD/SSD) that acts as virtual memory.

The Logic: Physical RAM is extremely fast but limited in capacity. When RAM fills up, the Linux kernel’s Memory Management Unit (MMU) identifies “inactive pages” (memory blocks not recently used) and moves them to the Swap space on the disk. This frees up high-speed RAM for active processes. Because disk I/O is vastly slower than RAM I/O, heavy reliance on swap degrades system performance (a state known as “thrashing”). Swap prevents out-of-memory (OOM) kernel crashes but is not a replacement for physical RAM.

Swappiness (The Control Valve)

Swappiness is a kernel parameter (0 to 100) defining how aggressively the system moves memory pages to Swap versus dropping the filesystem cache.
- Default: 60. A balanced approach.
- Low Value (e.g., 10): The kernel aggressively keeps application data in RAM and prefers dropping file cache. Ideal for interactive desktop systems where low latency is critical.
- High Value (e.g., 90): The kernel aggressively swaps out inactive memory to keep the file cache large. Ideal for batch-processing servers where high disk I/O throughput is more critical than instant memory access.
Commands
- To Check - cat /proc/sys/vm/swappiness OR sudo sysctl vm.swappiness
- To Change - sysctl vm.swappiness=20 → NOT RECOMMEND TO CHNAGE THE DEFAULT VALUE
  - Permanent: Add to /etc/sysctl.conf → vm.swappiness=20

Two Ways of Creating SWAP Space

1. Adding Swap via LVM - Partition or Volume [Recommended]

Use this when extending a dedicated swap partition backed by LVM.

Create LV: sudo lvcreate VolGroup00 -n LogVol02 -L 2G
Format as Swap: sudo mkswap /dev/VolGroup00/LogVol02
Persist (fstab): Add /dev/VolGroup00/LogVol02 swap swap defaults 0 0 to /etc/fstab
Reload Systemd: sudo systemctl daemon-reload
Activate: sudo swapon -v /dev/VolGroup00/LogVol02

2. Creating a Swap File

Use this for quick, flexible swap addition without altering partitions.

Allocate Space (e.g., 64MB): sudo dd if=/dev/zero of=/swapfile bs=1024 count=65536(Math: block size 1024 bytes * 65536 blocks = 64MB) OR bs=1M count=64
Format as Swap: sudo mkswap /swapfile
Secure File: sudo chmod 0600 /swapfile (Critical: Prevents non-root users from reading RAM data dumped to disk).
Persist (fstab): Add /swapfile swap swap defaults 0 0 to /etc/fstab
Reload Systemd: sudo systemctl daemon-reload
Activate: sudo swapon /swapfile

3. Removing a Swap File

Deactivate: sudo swapoff -v /swapfile
Remove Persistence: Delete the /swapfile line from /etc/fstab
Reload Systemd: sudo systemctl daemon-reload
Delete File: sudo rm /swapfile

Resize SWAP File
Section titled “Resize SWAP File”
- There is NO Direct Way to Increase or Decrease the Swap size.. but we can remove old one and create a new file with Desired Configuration..
- Eg: To increase from 2GB → 4GB:
  Terminal window
```
swapoff /swapfile
rm /swapfile
fallocate-l 4G /swapfile
chmod600 /swapfile
mkswap /swapfile
swapon /swapfile
```

Feature	Swap Partition	Swap File
Architecture	A dedicated block device/partition on the disk.	A dedicated file residing within an existing filesystem.
Performance	Historically faster (especially on HDDs due to contiguous placement reducing seek time).	Virtually identical in modern kernels (2.6+) as the kernel maps disk blocks directly, bypassing filesystem cache.
Flexibility	Rigid. Requires volume management (like LVM) or downtime to resize.	Highly flexible. Can be created, resized, or destroyed instantly on any mounted filesystem.
Recommendation	Red Hat recommends partitions for raw performance reliability.	Preferred for administrative ease and rapid scaling.