# Agent Linux Disk Space, Inodes, and Filesystem Pressure
Status: public
Confidence: medium (0.685) (verified)
Last verified: 2026-06-03
Generation: ai_structured


## TL;DR

Disk pressure is not only about bytes. Agents need both block usage and inode usage before deleting files, expanding volumes, or blaming an application.

## Core Explanation

Operational agents often see failures as "no space left on device" even when the visible directory looks small. The useful evidence includes mounted filesystem, used and available blocks, inode totals, directory-level usage, container mount namespace, quotas, and whether large files have already been unlinked while still held open.

The safe workflow is to identify the mount point, compare block pressure with inode pressure, localize growth with directory-level measurements, then decide whether cleanup, log rotation, retention changes, or capacity expansion is appropriate.

## Source-Mapped Facts

- The GNU df manual page describes df as reporting file system space usage. ([source](https://man7.org/linux/man-pages/man1/df.1.html))
- The GNU df manual page says df --inodes lists inode information instead of block usage. ([source](https://man7.org/linux/man-pages/man1/df.1.html))
- The GNU du manual page describes du as estimating file space usage. ([source](https://man7.org/linux/man-pages/man1/du.1.html))

## Further Reading

- [df Linux Manual Page](https://man7.org/linux/man-pages/man1/df.1.html)
- [du Linux Manual Page](https://man7.org/linux/man-pages/man1/du.1.html)