Disk Space and Disk Health

How full is it?

me@garden-laptop:~$ df -h

df reports each mounted file system’s size, used and available space, and use percentage; -h prints sizes like 48G instead of counting kilobytes ^{1 2}. df -h ~ shows just the file system your home folder is on. On Ubuntu you’ll see many small tmpfs file systems, which live in memory, and snap loop devices at 100%; the line that matters is usually /.

Which folders?

df says how full; du says what’s using it, by adding up folders ³:

me@garden-laptop:~$ du -sh ~/Videos
me@garden-laptop:~$ du -h -d 1 ~ | sort -h
me@garden-laptop:~$ sudo du -h -d 1 / 2>/dev/null | sort -h

• -s gives one total per argument; -h human sizes ³.
• -d 1 (--max-depth=1) lists each folder one level down ³.
• sort -h sorts human sizes properly, so the biggest come last (Shell course, module 2).

The last line surveys the whole system, with 2>/dev/null hiding the “permission denied” noise. Then drill down into whatever’s biggest. ncdu does the same interactively, with arrow keys to move into folders; install it with sudo apt install ncdu.

To find big single files:

me@garden-laptop:~$ find ~ -size +1G -exec ls -lh {} +

The usual suspects

Before deleting anything, know what it is. These are safe to clear:

• apt’s download cache. Every package apt downloads stays in /var/cache/apt/archives/ ⁴. sudo apt clean empties it; sudo apt autoremove removes dependencies nothing needs any more ⁵.
• The journal. journalctl --disk-usage, then sudo journalctl --vacuum-time=4weeks if it’s large (module 3) ⁶.
• Old snap revisions. snapd keeps previous revisions of each snap so you can revert; snap list --all shows them marked disabled, and the system option refresh.retain sets how many are kept ⁷.
• ~/.cache. By the freedesktop standard, programs keep “non-essential” cached data there ⁸. Deleting a program’s folder in it just makes the program rebuild it; close the program first.
• The Trash. Files deleted in the file manager go to ~/.local/share/Trash until you empty it.
• Your own files: old downloads, disk images (like that 6 GB Ubuntu .iso), videos.

Don’t delete things under /usr, /var/lib, or /boot by hand; let apt and the system manage those.

Is the drive healthy?

Drives watch themselves. SMART (Self-Monitoring, Analysis and Reporting Technology) is built into most hard disks and SSDs, and smartctl, from the smartmontools package, reads it ⁹:

me@garden-laptop:~$ sudo apt install smartmontools
me@garden-laptop:~$ sudo smartctl -H /dev/sda

-H prints the drive’s overall health. If it reports failing, the drive has either already failed or predicts its own failure within 24 hours: get your data off it as soon as you can ¹⁰. Use the disk name from lsblk, such as /dev/nvme0n1 on most laptops.

Drives can also test themselves. All the tests are safe for your data ⁹:

me@garden-laptop:~$ sudo smartctl -t short /dev/sda
me@garden-laptop:~$ sudo smartctl -l selftest /dev/sda
me@garden-laptop:~$ sudo smartctl -a /dev/sda

A short test takes a minute or two; a long test reads the whole surface and can take hours; -c shows how long each takes on your drive ⁹. -l selftest shows the results, and -a everything the drive reports ^{9 10}.

The Disks app shows the same thing without the terminal: select the drive, then the menu’s SMART Data & Self-Tests.

SSDs stay fast with TRIM, which tells the drive which blocks are free. Ubuntu enables periodic TRIM by default, through fstrim.timer, which runs weekly ¹¹; systemctl status fstrim.timer shows when it last ran.

The real protection

Health checks catch some failing drives early, but a drive can fail with no warning at all, and a SMART report can’t help with a deleted file, a stolen laptop, or a fire. Backups can. You have the pieces now: the snapshot script (Shell course), a timer that runs it nightly and catches up (module 3), and an ext4 drive mounted at the same place every time (this module). Every few months, restore a file from a snapshot to prove they work. A backup you’ve never restored from is a hope, not a backup.

Your turn

So

df -h shows how full each file system is; du -h -d 1 | sort -h, ncdu, and find -size show what’s using it. apt’s cache (apt clean), unneeded packages (apt autoremove), the journal (--vacuum-time), old snap revisions, ~/.cache, and the Trash are safe places to reclaim space. smartctl -H asks the drive about its own health, and a failing result means copy your data off now; short and long self-tests check further, and the Disks app shows the same. Ubuntu TRIMs SSDs weekly. None of it replaces backups you’ve tested.