Linux swappiness is a kernel parameter that defines how aggressively the system will use swap space relative to physical RAM. This setting lives under the hood of the kernel and directly influences memory management decisions, impacting application responsiveness, latency, and overall system stability. Understanding its behavior is essential for administrators tuning performance on workstations, servers, or cloud instances.
How Swappiness Works in the Kernel
The Linux kernel maintains a balance between keeping data in RAM and moving less active pages to disk. Swappiness is a tunable integer between 0 and 100 that adjusts the priority of swapping out inactive pages. A higher value makes the kernel more willing to evict processes from physical memory to swap, while a lower value keeps data in RAM as long as possible, reserving swap for emergency scenarios.
Interaction with Page Cache and Active Lists
Swappiness does not operate in isolation; it works alongside the page cache and the active/inactive LRU lists. The kernel first considers reclaiming reclaimable page cache before resorting to swapping application memory. With a low swappiness value, the kernel aggressively retains clean and hot file cache, reducing the need to write out anonymous memory. Conversely, a high swappiness setting encourages the kernel to push out anonymous pages sooner, which can free memory more aggressively but may increase disk I/O related to swap operations.
Default Values and Platform Differences
Most modern desktop distributions ship with a default swappiness of around 60, while server distributions often set it lower, such as 10 or even 1. These defaults reflect different priorities: desktops benefit from keeping applications in memory for snappy response, whereas servers aim to minimize swap I/O to protect throughput and predictability. The exact value can vary between Ubuntu, Debian, RHEL, and other distributions, so checking the current setting is always the first step in troubleshooting memory behavior.
Viewing the Current Swappiness Setting
You can inspect the current swappiness value by reading the relevant sysctl parameter. The runtime value is exposed through vm.swappiness in /proc/sys , and the same can be queried with the sysctl command. Persistent configuration is managed via /etc/sysctl.conf or files in /etc/sysctl.d/ , ensuring the setting survives reboots across restarts and migrations.
Tuning Swappiness for Performance and Latency
Adjusting swappiness is a common optimization technique, but the outcome depends heavily on workload characteristics. For latency-sensitive applications like databases or real-time services, lowering swappiness reduces the risk of sudden swap-induced pauses. On the other hand, memory-heavy batch jobs on machines with ample RAM and slow disks might benefit from a higher value to proactively free memory and avoid out-of-memory kills. The key is to align the setting with the expected memory pressure and I/O profile.
Practical Steps to Change Swappiness
To adjust swappiness temporarily, you can write a new value to /proc/sys/vm/swappiness using sysctl or direct echo commands. For a permanent change, add or modify vm.swappiness=10 in the sysctl configuration, then reload the settings. It is wise to monitor metrics such as swap usage, major faults, and disk I/O after the change, validating that the new behavior matches the intended performance goals without introducing regressions.
