Working with files and folders in Linux

You’ll build a small practice folder in your home directory on rocky01, fill it with files, and then copy, move, link and delete them. Everything stays inside ~/practice, so nothing on the system is at risk. Log in as student over SSH from your laptop (ssh -p 2211 student@127.0.0.1).

Your inode numbers, times and sizes will differ from the ones below. On ubu01 you won’t see the dot after the permissions (no SELinux), directories show 4096 as their size on ext4, and stat has no Context line.

1Know where you are

Run this:

cd
pwd
cd /etc
pwd
cd -

You should see this:

/home/student
/etc
/home/student

cd with nothing after it takes you home, and cd - jumps back to where you were and prints the folder. Get used to running pwd before any delete: most “wrong folder” accidents happen because someone assumed where they were.

✅ Checkpoint: pwd prints /home/student after cd -.

2Make a folder tree in one go

Run this:

mkdir practice/reports/2026/october
mkdir -p practice/reports/2026/october
ls -R practice

You should see this:

mkdir: cannot create directory 'practice/reports/2026/october': No such file or directory
practice:
reports

practice/reports:
2026

practice/reports/2026:
october

practice/reports/2026/october:

Plain mkdir refuses because the parent folders don’t exist yet. mkdir -p creates every missing parent and stays quiet if the folder is already there, which is why scripts use it. ls -R lists the tree recursively.

✅ Checkpoint: ls -R shows practice, reports, 2026 and october.

3Create files and read ls -l

Run this:

cd ~/practice
touch notes.txt app.log app.log.1 app.log.2 report1.csv report2.csv report10.csv
printf 'web01\nweb02\ndb01\n' > servers.txt
ls -l

You should see something like this:

total 4
-rw-rw-r--. 1 student student  0 Oct 27 10:16 app.log
-rw-rw-r--. 1 student student  0 Oct 27 10:16 app.log.1
-rw-rw-r--. 1 student student  0 Oct 27 10:16 app.log.2
-rw-rw-r--. 1 student student  0 Oct 27 10:16 notes.txt
-rw-rw-r--. 1 student student  0 Oct 27 10:16 report1.csv
-rw-rw-r--. 1 student student  0 Oct 27 10:16 report10.csv
-rw-rw-r--. 1 student student  0 Oct 27 10:16 report2.csv
drwxrwxr-x. 3 student student 18 Oct 27 10:15 reports
-rw-rw-r--. 1 student student 17 Oct 27 10:16 servers.txt

touch made seven empty files (size 0). printf wrote three lines into servers.txt, which is 17 bytes including the line endings; Lesson 13 explains the > sign. The files are rw-rw-r-- because student’s umask is 0002 on both Rocky and Ubuntu (user private groups), and reports shows a size of 18 because a small XFS directory stores its entries inside the inode.

✅ Checkpoint: You can point at the type, permissions, link count, owner, size, time and name of servers.txt.

4Let the shell pick files with wildcards

Run this:

ls app.log*
ls report?.csv
ls report[0-9]*.csv
echo report*.csv
ls *.CSV

You should see this:

app.log
app.log.1
app.log.2
report1.csv
report2.csv
report1.csv
report10.csv
report2.csv
report1.csv report10.csv report2.csv
ls: cannot access '*.CSV': No such file or directory

report?.csv skips report10.csv because ? stands for one character only. echo shows you the list bash built, which is the safest way to test a pattern before you hand it to rm. The last error mentions *.CSV literally because nothing matched (names are case-sensitive), so bash passed the pattern through untouched.

✅ Checkpoint: echo report*.csv prints three names and you can say why report?.csv printed only two.

5Copy safely

Run this (answer n to the question):

cp servers.txt servers.bak
cp -i servers.txt servers.bak
cp reports reports-backup
cp -r reports reports-backup
ls -d reports*

You should see this:

cp: overwrite 'servers.bak'? n
cp: -r not specified; omitting directory 'reports'
reports
reports-backup

The first copy overwrites silently if servers.bak exists, which is the default and the danger. -i asks first. A folder needs -r, and when you back up config at work, use cp -a (or cp -p) so the owner, permissions and times stay the same as the original.

✅ Checkpoint: ls -d reports* shows both reports and reports-backup.

6Rename with mv and watch the inode

Run this:

ls -i notes.txt
mv notes.txt todo.txt
ls -i todo.txt

You should see something like this:

33575013 notes.txt
33575013 todo.txt

The inode number didn’t change. mv inside one file system rewrites the directory entry and leaves the data where it is. Across file systems (say from /home to /dev/shm) it has to copy and then delete, which takes longer and gives the file a new inode.

✅ Checkpoint: The inode number is the same before and after the rename.

7Delete with care

Run this (answer y to the last question):

mkdir -p old/stuff
rmdir old
rm old
rm -i app.log.2
rm -r reports-backup old
ls

You should see this, then the remaining files:

rmdir: failed to remove 'old': Directory not empty
rm: cannot remove 'old': Is a directory
rm: remove regular empty file 'app.log.2'? y

rmdir only removes empty folders, plain rm refuses folders, and rm -r removes a folder and everything in it. There is no recycle bin on a Linux server, so a deleted file is gone unless a backup or snapshot has it.

✅ Checkpoint: old, reports-backup and app.log.2 are gone; todo.txt and servers.txt are still there.

8Look inside the inode with stat

Run this:

stat servers.txt

You should see something like this:

  File: servers.txt
  Size: 17        	Blocks: 8          IO Block: 4096   regular file
Device: 253,0	Inode: 33575012    Links: 1
Access: (0664/-rw-rw-r--)  Uid: ( 1000/ student)   Gid: ( 1000/ student)
Context: unconfined_u:object_r:user_home_t:s0
Access: 2026-10-27 10:16:40.512334018 +0530
Modify: 2026-10-27 10:16:40.512334018 +0530
Change: 2026-10-27 10:16:40.512334018 +0530
 Birth: 2026-10-27 10:16:40.512334018 +0530

This is the society register entry: inode number, link count, permissions in number and letter form, owner and group by UID and GID, the SELinux label (Rocky only) and four times. Modify changes when the content changes, Change when the inode changes (chmod, chown, a new hard link), Access when it’s read (Linux updates it lazily), and Birth is when it was created. To get one field in a script, ask for it: stat -c %i servers.txt prints only the inode number.

✅ Checkpoint: You can read the inode number and the link count (1) from stat.

9Make a hard link and a soft link

Run this:

ln servers.txt servers-hard.txt
ln -s servers.txt servers-soft.txt
ls -li servers*.txt

You should see something like this:

33575012 -rw-rw-r--. 2 student student 17 Oct 27 10:16 servers-hard.txt
33575021 lrwxrwxrwx. 1 student student 11 Oct 27 10:24 servers-soft.txt -> servers.txt
33575012 -rw-rw-r--. 2 student student 17 Oct 27 10:16 servers.txt

servers.txt and servers-hard.txt share inode 33575012 and the link count went up to 2. The soft link has its own inode, type l, and its size is 11 because it stores the 11 characters “servers.txt”.

Now remove the original name and read both links:

rm servers.txt
ls -l servers-hard.txt servers-soft.txt
cat servers-hard.txt
cat servers-soft.txt

You should see something like this:

-rw-rw-r--. 1 student student 17 Oct 27 10:16 servers-hard.txt
lrwxrwxrwx. 1 student student 11 Oct 27 10:24 servers-soft.txt -> servers.txt
web01
web02
db01
cat: servers-soft.txt: No such file or directory

The hard link still has the data, because the inode had two names and only one was removed. The soft link is now broken: it points at a name that no longer exists. Put the file back so the next lesson has it:

mv servers-hard.txt servers.txt

✅ Checkpoint: cat servers-soft.txt works again after the mv, because the name it points to is back.