I'm trying to compress a directory with bzip2. According to documentation I have to 'bzip2 whatever.bz2 dirtopack'. But it gives me a big:

What the hell am I doing wrong? I've tried some switches but nothing works. Tar works fine but bzip2 is more powerful.

You seem to misunderstand the difference between tar and bzip2.

tar is a file archive utility that takes a number of files, with pathnames, and puts them into a single binary file. It doesn't do anything to compress the data.

bzip2 (like gzip and compress) is a file (or stream) compression utility, that takes a single file and makes it smaller.

If you want to create an archive of a directory, you first need to create a tar archive, then bzip2 compress the archive; this is why “tarball”s often have the extension .tar.bz2You can also do this in one step:This compresses better than all-in-one formats like ZIP, which compress each file seperately then append then put the compressed files into an archive.

If you want to compress each file within a directory, you can use To do this recursively you'll need to use find:

1 members found this post helpful.

I was googling how to do the best compression of a directory on linux and I found that if you have memory XZ is the best as they explain it here how to compress a directory in linux

I always just use the 'j' switch with 'tar'.

It all depends on what you're trying to do, really.

gzip (-z with tar) is generally fastest to compress and uncompress, but only provides limited compression.
bzip (-j with tar) provides slightly better compression than gzip (except for very small files - unlikely with tar) and takes a little more cpu and memory, but is generally a good alternative.
xz (-J with tar) provides the best compression ratio, but is expensive in terms of time and memory used, so generally best for archiving or creating big distributions over slower networks.

1 members found this post helpful.