cpio, tar, gzip, bzip2, xz

cpio

Warning: cpio can overwrite the files of your system. The author is not responsible of any damage to the operating system.

A couple of examples on how to create a cpio file:

$ find . | cpio -ov > archive.cpio

$ ls file? | cpio -ov > archive2.cpio

Where:

-o means output

-v means verbose

Notice that cpio my use the absolute path or the relative path of a file, depending on what was given to it. When using the absolute path there is the risk of overwriting an important file, for example, /etc/passwd.

• Older Files on Disk: If the file on your computer happens to be older than the one in the archive (for example, if you're restoring a config file you recently accidentally broke), cpio will overwrite it by default because it assumes you want the "newer" version from the archive.

• The --no-absolute-filenames Requirement: You have to force the command to remove those paths to be safe. If an archive has absolute paths and you forget that specific flag, you lose control over where those files land.

To get the files from the cpio file:

$ cpio -itv < archive.cpio # It will only list the contents of the cpio file

-t means table of contents (list only, doesn't actually extract the contents)

-i input

-v verbose

To actually extract your files from the cpio file:

$ cpio -iv < archive.cpio

But remember to avoid the absolute path with:

$ cpio -iv --no-absolute-filenames < archive.cpio

-d: Directories (optional here, but recommended if your archive contains subfolders so cpio can recreate them). So it would be -idv flags instead.

tar

Combine archiving (tar) with compression (gz):

$ tar -zcvf archive.tar.gz myfiles/ # '-z' is the gzip option

Only showing the contents without extracting the files:

$ tar -tvf archive.tar.gz

Extracting files (with a relative path):

$ tar -xvf archive.tar.gz

Options to create/unpack tarballs:
-z --gzip
-j --bzip2
-J --xz

That's how you name compressed files:

.tar.gz or .tgz
.tar.bz or .tbz
.tar.xz or .txz

These are the different flags for just tar files (without compression):

tar -cvf

tar -tf

tar -xvf

These are the three different compression modes:

tar -zcvf (gzip)

tar -jcvf (bzip2)

tar -Jcvf (xz)

tar vs cpio

1. Use tar when...

• Creating manual backups: It is much easier to use because you can just point it at a directory (e.g., tar -cvf backup.tar /home/user).

• Sending files to others: It is the universal standard on Linux/Unix systems.

• Handling compression easily: tar has built-in flags like -z (gzip) or -j (bzip2), so you don't need pipes.

2. Use cpio when...

• Dealing with "Special" files: cpio is famous for its ability to handle device files (like those in /dev) and hard links more faithfully than older versions of tar.

• Creating initramfs (Boot images): This is the most common use today. The Linux kernel uses cpio to load the temporary file system used during the boot process.

• Complex filtering: Because cpio takes a list of filenames from find, you can be extremely specific about what gets archived.

  • Example: Archiving only files modified in the last 2 days that are larger than 1MB: find . -mtime -2 -size +1M | cpio -ov > specific_backup.cpio

gzip / bzip2 / xz

If you find compressed files, you can decompress them with:

gunzip myfile.txt.gz or bunzip2 myfile.txt.bz2 or unxz myfile.txt.xz

What would this be the output for the following commands? A single file or multiple files?:

$ touch myfile1 myfile2 myfile3

$ ls myfile?

$ ls myfile? | xargs gzip

The answer is: multiple files.

$ ls myfile*

myfile1.gz myfile2.gz myfile3.gz

If you want to wrap several files into one file, you need to use tar.

To see the contents of a compressed file without decompressing it, you can use:

gzip -> zcat

bzip2 -> bzcat

xz -> xzcat

For example, try this: