So you are wondering, how to check disk space in Linux from the command line? This guide shows you exactly that, using the two tools you will reach for 90% of the time, df (to check free space) and du (to check disk usage by folder), plus the commands that tell you what is eating your storage. Every command below was run on a live Ubuntu server and the output you see is the real, untouched result, including a root filesystem that is sitting at a worrying 93% full.
In Linux, checking disk space is a routine but critical task. You either want a quick overview of how much free space is left on each partition, or you are responding to a No space left on device error and need to find what is using up the disk. We will cover both, from a fast one-liner to a full investigation.
Quick reference: Linux disk space commands
| Goal | Command |
|---|---|
| Check free disk space (all filesystems) | df -h |
| Check disk space with filesystem type | df -hT |
| Check space on one mount point | df -h / |
| Show real disks only (hide tmpfs) | df -h -x tmpfs -x devtmpfs |
| Check inode usage (hidden "disk full") | df -i |
| Disk usage of a folder (total) | du -sh /path |
| Size of each subdirectory, sorted | du -h --max-depth=1 /path | sort -rh |
| Biggest directories on one filesystem | du -xh --max-depth=1 /mount | sort -rh |
| Find largest files | find /path -xdev -type f -size +500M |
| Device and partition layout | lsblk -e7 |
| Size of systemd journal logs | journalctl --disk-usage |
| Physical disk size (hardware) | sudo fdisk -l |
1. df command – check free and used disk space
The df command (short for disk free) is the first tool to reach for when you want to check disk space in Linux. It reports how much space is total, used and available on every mounted filesystem. The -h (human-readable) flag prints sizes in K, M and G instead of raw 1‑kilobyte blocks:
df -hFilesystem Size Used Avail Use% Mounted on
tmpfs 847M 3.0M 844M 1% /run
/dev/mapper/ubuntu--vg-ubuntu--lv 43G 38G 3.0G 93% /
tmpfs 4.2G 0 4.2G 0% /dev/shm
tmpfs 5.0M 8.0K 5.0M 1% /run/lock
tmpfs 4.2G 32M 4.2G 1% /tmp
/dev/sda2 2.0G 222M 1.6G 13% /boot
tmpfs 847M 204K 847M 1% /run/user/1000The columns are: filesystem device, total Size, Used space, Available space, Use% (percentage used) and the Mounted on point. Here the root filesystem / is at 93% used with only 3.0G free, which is the partition we would investigate.
Show the filesystem type with df -hT
Add -T to print the filesystem type (ext4, xfs, tmpfs, etc.). This is useful to tell real disks apart from in-memory tmpfs mounts:
df -hTFilesystem Type Size Used Avail Use% Mounted on
tmpfs tmpfs 847M 3.0M 844M 1% /run
/dev/mapper/ubuntu--vg-ubuntu--lv ext4 43G 38G 3.0G 93% /
tmpfs tmpfs 4.2G 0 4.2G 0% /dev/shm
/dev/sda2 ext4 2.0G 222M 1.6G 13% /boot
/dev/sr0 iso9660 51M 51M 0 100% /media/golinuxcloud/VBox_GAs_7.2.01Show only real disks (hide tmpfs noise)
On modern systems the df output is cluttered with tmpfs pseudo-filesystems. Use -x to exclude one or more types so you see only the physical disks and partitions:
df -h -x tmpfs -x devtmpfsFilesystem Size Used Avail Use% Mounted on
/dev/mapper/ubuntu--vg-ubuntu--lv 43G 38G 3.0G 93% /
/dev/sda2 2.0G 222M 1.6G 13% /boot
/dev/sr0 51M 51M 0 100% /media/golinuxcloud/VBox_GAs_7.2.01Check disk space on a single filesystem
If you only care about one mount point, pass its path. df -h / removes all guesswork about which line is the root filesystem:
df -h /Filesystem Size Used Avail Use% Mounted on
/dev/mapper/ubuntu--vg-ubuntu--lv 43G 38G 3.0G 93% /Get a grand total across disks
Add --total to append a summary row that sums all listed filesystems, handy for a single "how much storage do I have left" number:
df -h --total -x tmpfs -x devtmpfsFilesystem Size Used Avail Use% Mounted on
/dev/mapper/ubuntu--vg-ubuntu--lv 43G 38G 3.0G 93% /
/dev/sda2 2.0G 222M 1.6G 13% /boot
/dev/sr0 51M 51M 0 100% /media/golinuxcloud/VBox_GAs_7.2.01
total 45G 38G 4.6G 90% -Check inode usage with df -i
A filesystem can report No space left on device even when df -h shows free space, because it has run out of inodes (one inode is consumed per file, directory or symlink). Use -i to check this, and watch the IUse% column:
df -iFilesystem Inodes IUsed IFree IUse% Mounted on
/dev/mapper/ubuntu--vg-ubuntu--lv 2818048 712525 2105523 26% /
tmpfs 1083776 6 1083770 1% /run/lock
/dev/sda2 131072 330 130742 1% /bootHere the root filesystem uses only 26% of its inodes, so inode exhaustion is not the problem. If IUse% were at 100%, you would be out of inodes regardless of free bytes. You can see the full set of options on the man page of df command.
2. du command – check disk usage by folder
Once df tells you which filesystem is full, the du command (short for disk usage) tells you what is filling it. du estimates the space used by files and directories, giving you the flexibility to check disk space by folder in Linux.
Total size of a single folder
The most common use is du -sh, where -s summarizes to a single total and -h makes it human-readable:
sudo du -sh /var/log313M /var/logSize of each subdirectory, largest first
To find the heaviest subdirectory, list one level deep with --max-depth=1 and pipe it to sort -rh (the -h flag of sort understands human-readable sizes like 2K, 4M, 1G, and -r reverses to largest-first):
sudo du -h --max-depth=1 /var/log | sort -rh | head313M /var/log
297M /var/log/journal
2.0M /var/log/sysstat
1.1M /var/log/installer
320K /var/log/apt
84K /var/log/unattended-upgrades
36K /var/log/cupsThe result is immediate: of the 313M in /var/log, the systemd journal directory alone accounts for 297M, so that is where the cleanup effort should focus.
Show only files above a size threshold
du lists everything by default. Add --threshold (or -t) to print only entries larger than a given size, which is perfect for spotting big files quickly:
sudo du -ah --threshold=1M /var/log | sort -rh | head313M /var/log
297M /var/log/journal/22e961a5781c41b3bed7f02bb85a0a9b
297M /var/log/journal
8.1M /var/log/journal/.../system.journal
8.1M /var/log/journal/.../user-1000@...journal~You can find more options on the man page of du command. If you want to manipulate the output further, our guide on awk pairs well with du for custom reports.
3. Find what is eating your disk space (the df → du → find workflow)
When a server reports No space left on device, the reliable workflow is df → du → find: find the full filesystem, find the biggest directory on it, then find the biggest files. Staying on one filesystem matters, so use du -x to stop it from crossing into other mounts.
Step 1 – find the biggest top-level directories on the root filesystem:
sudo du -xh --max-depth=1 / | sort -rh | head38G /
23G /var
6.9G /home
6.2G /usr
489M /root
15M /opt
13M /etc/var is the clear culprit at 23G. You would then drill into it (sudo du -xh --max-depth=1 /var | sort -rh) and keep following the biggest branch.
Step 2 – list the largest individual files with find. The -xdev flag keeps the search on a single filesystem and -size +200M filters to files larger than 200 MB:
sudo find / -xdev -type f -size +200M -printf '%s\t%p\n' | sort -rn | head1073741824 /swap.img
724512768 /home/golinuxcloud/golinuxcloud-static/.git/objects/pack/pack-b78bd4f8....pack
635518976 /var/lib/snapd/snaps/gnome-46-2404_153.snap
541278208 /var/lib/snapd/snaps/gnome-42-2204_226.snap
414167040 /var/lib/snapd/snaps/mesa-2404_1165.snap
294649856 /var/lib/containerd/.../blobs/sha256/c6c3054e9f3b...Now the heavy hitters are obvious: a 1 GB swap file, a large Git pack, cached snap packages and container blobs. Cleanup becomes a deliberate decision instead of guesswork.
Interactive alternative: ncdu
For an interactive, navigable view of disk usage, the ncdu (NCurses Disk Usage) tool is a favourite. It is not installed by default, so add it first:
# Debian/Ubuntu
sudo apt install ncdu
# RHEL/Rocky/Fedora
sudo dnf install ncduThen run ncdu / and browse the tree with the arrow keys, deleting files with d. It is the easiest way to explore which directories are consuming the most space without memorising flags.
4. Check the systemd journal size
On systemd distributions the binary journal under /var/log/journal is one of the most common reasons a root filesystem quietly fills up, exactly as we saw above. Check its size directly with:
journalctl --disk-usageArchived and active journals take up 296.1M in the file system.If it is large, cap it with sudo journalctl --vacuum-size=200M or sudo journalctl --vacuum-time=2weeks. For more on reading and filtering these logs, see our guide on journalctl.
5. lsblk command – device and partition layout
lsblk lists block devices in a tree, which is the clearest way to understand which partitions and logical volumes exist and where they are mounted. On a desktop the output is buried under dozens of snap loop devices, so use -e7 to exclude them (major number 7 = loop devices):
lsblk -e7NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
sda 8:0 0 25G 0 disk
├─sda1 8:1 0 1M 0 part
├─sda2 8:2 0 2G 0 part /boot
└─sda3 8:3 0 23G 0 part
└─ubuntu--vg-ubuntu--lv 252:0 0 43G 0 lvm /
sdb 8:16 0 20G 0 disk
└─ubuntu--vg-ubuntu--lv 252:0 0 43G 0 lvm /
sdc 8:32 0 15.7G 0 disk
sr0 11:0 1 50.7M 0 rom /media/golinuxcloud/VBox_GAs_7.2.01This system has three disks (sda, sdb, sdc), and the root LVM volume is built across sda3 and sdb. You can also request custom columns including filesystem usage:
lsblk -e7 -o NAME,SIZE,FSTYPE,FSAVAIL,FSUSE%,MOUNTPOINTNAME SIZE FSTYPE FSAVAIL FSUSE% MOUNTPOINT
sda 25G
├─sda1 1M
├─sda2 2G ext4 1.6G 11% /boot
└─sda3 23G LVM2_member
└─ubuntu--vg-ubuntu--lv 43G ext4 3G 88% /
sdb 20G LVM2_member
sdc 15.7G
sr0 50.7M iso9660 0 100% /media/golinuxcloud/VBox_GAs_7.2.01To confirm where a path is mounted and its source device, findmnt is the cleanest option:
findmnt /TARGET SOURCE FSTYPE OPTIONS
/ /dev/mapper/ubuntu--vg-ubuntu--lv ext4 rw,relatimeFor more ways to inspect mounted filesystems, see how to check mounted filesystems.
6. Check the physical disk size (hardware capacity)
The commands above report filesystem space. If instead you want the raw size of the physical disks attached to the machine, use fdisk -l and filter to the real /dev/sd* disks:
sudo fdisk -l | grep -E "^Disk /dev/sd"Disk /dev/sda: 25 GiB, 26843545600 bytes, 52428800 sectors
Disk /dev/sdb: 20 GiB, 21474836480 bytes, 41943040 sectors
Disk /dev/sdc: 15.7 GiB, 16861716480 bytes, 32933040 sectorsfdisk is primarily for creating and modifying partitions, but -l is a quick way to confirm exact disk capacity. The parted command (sudo parted -l) and lsblk shown earlier give the same hardware view. To understand the different disk and interface types you might see, refer to our overview of disk types in Linux.
7. Troubleshooting: "No space left on device" with free space showing
This is the question that drives most disk-space searches, so here are the four real causes and how to confirm each one.
Cause 1 – Out of inodes. Run df -i; if IUse% is 100% you are out of inodes (usually millions of tiny cache or session files) even though df -h shows free bytes. Delete or archive the small files to recover.
Cause 2 – Deleted files still held open. If a process is still writing to a file you deleted, the kernel will not release the space until the process closes it. Find these with:
sudo lsof +L1COMMAND PID USER FD TYPE DEVICE SIZE/OFF NLINK NODE NAME
gnome-she 7581 golinuxcloud 39u REG 0,1 68403 0 1092 /memfd:mutter-shared (deleted)The NLINK column at 0 and the (deleted) marker identify the file, and SIZE/OFF shows the space it is still holding. Restart the owning process, or reclaim the space without restarting by truncating the file descriptor: : > /proc/PID/fd/FD.
Cause 3 – Reserved blocks. ext2/3/4 filesystems reserve ~5% of space for root so the system stays usable when full. A normal user can hit "disk full" while that reserve is intact. Confirm it with:
sudo tune2fs -l /dev/mapper/ubuntu--vg-ubuntu--lv | grep -iE "block count|reserved block count"Block count: 11270144
Reserved block count: 511041That is 511041 of 11270144 blocks (about 4.5%, ~2 GB at 4K blocks) reserved for root. You can lower it on data volumes with sudo tune2fs -m 1 /dev/....
Cause 4 – Files hidden under a mount point. If data was written to a directory and another filesystem was later mounted on top, the original files still consume space but are invisible. Compare df against du, or unmount suspect mounts and check what was underneath.
Summary
There are several ways to check disk space in Linux, but two commands do most of the work. Use the df command (df -h) to check free and used disk space across all mounted filesystems, and df -h / or df -hT when you want a single mount point or the filesystem type. Use the du command (du -sh /path and du -h --max-depth=1 /path | sort -rh) to check disk usage by folder and pinpoint what is consuming storage.
When a disk is actually full, follow the df → du → find workflow to move from the full filesystem to the biggest directory to the largest files. And when Linux reports No space left on device while still showing free space, check inodes with df -i, look for deleted-but-open files with lsof +L1, and remember the ext4 reserved-blocks pool. Layer in lsblk, findmnt and journalctl --disk-usage for a complete picture, and you can diagnose any disk space issue from the Linux command line with confidence.
