The df command is one of the most frequently used Linux command line utilities for checking disk space usage. As a Linux system administrator, having an in-depth understanding of df is critical for managing storage infrastructure.
This comprehensive guide will explore all facets of df usage – from basic syntax to advanced troubleshooting.
How Does df Work?
The df command provides a summary of unused and used disk space on all mounted filesystems in Linux by parsing two key sources of information:
/etc/mtab
This file contains a list of currently mounted filesystems. df relies on /etc/mtab to identify which disks/partitions to probe for statistics.
/sys/ Filesystems
This virtual filesystem provides detailed metrics on disk usage for each filesystem on a running Linux system. df queries files within /sys/ to build its output report.
So in summary, df combines current mount information with disk statistics pulled from /sys to produce a storage usage summary.
Display Disk Space for Mounted Disks
The most basic df invocation provides disk space statistics for mounted filesystems only:
$ df
Filesystem 1K-blocks Used Available Use% Mounted on
udev 8004592 0 8004592 0% /dev
tmpfs 1628420 1948 1626472 1% /run
/dev/vda1 825564424 66011156 71436940 9% /
tmpfs 8142088 4 8142084 1% /dev/shm
tmpfs 5120 4 5116 1% /run/lock
tmpfs 8142088 0 8142088 0% /sys/fs/cgroup
/dev/vda15 1018084 56720 892456 6% /boot
tmpfs 1628416 4 1628412 1% /run/user/0This output displays storage usage clearly showing the filesystem device name, 1K-block totals (KB size), used vs available blocks, percentage used and where each filesystem is mounted.
Show All Filesystems with -a
The -a flag tells df to show information on all filesystems, even if not mounted:
$ df -a
Filesystem 1K-blocks Used Available Use% Mounted on
udev 8004592 0 8004592 0% /dev
tmpfs 1628420 1948 1626472 1% /run
/dev/vda1 825564424 66011156 71436940 9% /
tmpfs 8142088 4 8142084 1% /dev/shm
tmpfs 5120 4 5116 1% /run/lock
tmpfs 8142088 0 8142088 0% /sys/fs/cgroup
/dev/vda15 1018084 56720 892456 6% /boot
tmpfs 1628416 4 1628412 1% /run/user/0This can help identify disk partitions that are unmounted or offline. The output includes dummy and duplicate filesystems also.
Show Totals with –total
Adding the –total flag provides a summary line showing combined totals for all disks:
$ df --total
Filesystem 1K-blocks Used Available Use% Mounted on
udev 8004592 0 8004592 0% /dev
tmpfs 1628420 1948 1626472 1% /run
/dev/vda1 825564424 66011156 71436940 9% /
tmpfs 8142088 4 8142084 1% /dev/shm
tmpfs 5120 4 5116 1% /run/lock
tmpfs 8142088 0 8142088 0% /sys/fs/cgroup
/dev/vda15 1018084 56720 892456 6% /boot
tmpfs 1628416 4 1628412 1% /run/user/0
total 843573232 67387840 71644176 9% /The total summary quantifies disk space and utilization for all mounted filesystems.
Show Human-Readable Sizes with -h
The -h flag displays disk space in human-readable format (KB, MB, GB), making it easier to parse:
$ df -h
Filesystem Size Used Avail Use% Mounted on
udev 7.7G 0 7.7G 0% /dev
tmpfs 1.6G 1.9M 1.6G 1% /run
/dev/vda1 79G 64G 15G 81% /
tmpfs 7.9G 4.0K 7.9G 1% /dev/shm
tmpfs 5.0M 4.0K 5.0M 1% /run/lock
tmpfs 7.9G 0 7.9G 0% /sys/fs/cgroup
/dev/vda15 996M 554M 442M 56% /boot
tmpfs 1.6G 4.0K 1.6G 1% /run/user/0Human sizes make it easier to process disk usage at a glance.
Select Specific Units with kb, mb, gb
To specify the units for disk space amounts, use kb, mb or gb arguments:
$ df -k
$ df -m
$ df -g For example, to show results in megabytes:
$ df -m
Filesystem 1M-blocks Used Available Use% Mounted on
udev 78408 0 78408 0% /dev
tmpfs 15948 14 15935 1% /run
/dev/vda1 8078716 25244 758252 1% /
tmpfs 79362 12 79350 1% /dev/shm
tmpfs 14 12 14 0% /run/lock
tmpfs 79362 0 79362 0% /sys/fs/cgroup
/dev/vda15 988 552 436 56% /boot
tmpfs 15948 4 15943 1% /run/user/0Analyze Specific Filesystem Types
The -t flag shows information for only a specified filesystem type, as denoted by /etc/fstab, for example:
$ df -t ext4
$ df -t xfsThis can help analyze disk space specifically on the required filesystems.
Exclude Specified Filesystems
To omit certain filesystems from the df output, use the -x flag:
$ df -x tmpfsThis is useful for focusing your disk report only on physical storage volumes.
Show Inode Information with -i
The -i flag displays additional details about inodes – the underlying filesystem structure that manages individual files:
$ df -iFilesystem Inodes IUsed IFree IUse% Mounted on
udev 2006148 561 2005587 1% /dev
tmpfs 2035604 1490 2034114 1% /run
/dev/vda1 52426752 3787393 48639589 8% /
tmpfs 2035604 19 2035585 1% /dev/shm
tmpfs 2035604 5 2035599 1% /run/lock
tmpfs 2035604 2 2035602 1% /sys/fs/cgroup
/dev/vda15 261273 7150 254123 3% /boot
tmpfs 2035604 20 2035584 1% /run/user/0Monitoring inodes helps identify potential inode depletion issues.
Troubleshooting df Errors
Some common errors with df and their usual causes:
Stale File Handle – Filesystem was unmounted/removed
Transport Endpoint Not Connected – Network fileshare unavailable
Timeout Acquiring Fileset Information – Underlying storage is inaccessible
Debug tips for above errors:
- Check filesystem mounts with mount command
- Verify disk connectivity and configuration
- Test network connectivity to remote filesystems
Best Practices for Using df
To leverage df most effectively:
- Set up thresholds (% used) for alerting
- Schedule periodic df checks with cron
- Parse df output to feed monitoring tools
- Baseline normal usage with historical df data
- Script df to run checks autonomously
For example, a simple baseline monitoring script:
#!/bin/bash
baseline=/var/opt/diskspace/baseline.txt
df -h > tempdiskuse.txt
percentused=$(sed -n ‘s/\S*% \/$/\1/p‘ tempdiskuse.txt)
echo "$(date) percentused = $percentused" >> $baseline
if [[ $percentused > 70 ]]; then
echo "Alert: Disk space crossed 70%" | mail -s "Disk Space Alert" admin@example.com
fi This runs df periodically, storing usage % over time and alerting when thresholds crossed.
Visualizing Disk Usage
Charts and data visualizations are useful for identifying disk usage growth patterns over longer durations:
Tools like pydf, gdmap and durep can be used to generate graphs from df statistics.
Comparing df to Related Tools
du – disk usage summary per directory rather than whole filesystem
find – find files matching criteria rather than overall usage
ncdu – interactive ncurses-based UI for drilling into disk usage
pydf – Python module to plot interactive graphs from df stats
So while df provides a filesystem/disk level overview, utilities like du and ncdu complement it by letting you analyze disk space usage at finer granularity.
Under the Hood: df Implementations
Modern Linux distributions implement df via calls to:
/proc/self/mountinfo – gathers mounted filesystem data
/sys/block filestat data – provides underlying device usage statistics
Earlier Linux systems relied instead on:
/etc/mtab entries – to get mountpoints
sys_statvfs system call – pull filesystem metrics
So while df output remains consistent in all Linux variants, the back-end implementation differs. Similarly df functionality varies in other UNIX systems – FreeBSD, OpenBSD, Solaris, HP-UX etc. But POSIX standards ensure baseline compatibility.
Summary
The humble but essential df command offers valuable insights into disk usage across any Linux environment. Combining its capabilities for customizable output with monitoring best practices allows effectively managing storage growth and preventing outages. This guide has covered df from basic to advanced usage – enabling you to squeeze the most value from this ubiquitous Linux administrative tool.