Excluding files from a tar archive

I was trying to make a backup, but exclude my working directory, which just has a lot of temp files in it.

. . . easier said than done . . .

The concept is pretty easy, simply use the --exclude="file2exclude" switch option in the command, and the file(s)/directory(ies) listed will be excluded.

It didn't work.

Doing some digging into it, I found various posts about where to place the --exclude switch.

The tl;dr version is to put the --exclude between the archive location, and what is being added to the archive.

• I'm running Ubuntu 20.04, which comes with tar version 1.3.0:

  tar --version
  tar (GNU tar) 1.30
  Copyright (C) 2017 Free Software Foundation, Inc.
  License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>.
  This is free software: you are free to change and redistribute it.
  There is NO WARRANTY, to the extent permitted by law.
  
  Written by John Gilmore and Jay Fenlason.
  

• In this example, I have a file structure that looks like this:

  tree
  .
  ├── contact.html
  ├── db.bkup.tar.gz
  ├── index.html
  ├── internal
  │   ├── import.db
  │   └── scratchfile.txt
  ├── passwordfile.txt
  └── public
  
  2 directories, 6 files
  

  I want to exclude the internal directory, and also the passwordfile.txt file from the archive.

• Some posts said to put the --exclude switch at the end of the tar command, but that didn't work:

  tar cfvz ../backup.tar.gz * --exclude="internal" --exclude="passwordfile.txt"
  contact.html
  db.bkup.tar.gz
  index.html
  internal/
  internal/import.db
  internal/scratchfile.txt
  passwordfile.txt
  public/
  tar: The following options were used after any non-optional arguments in archive create or update mode.  These options are positional and affect only arguments that follow them.  Please, rearrange them properly.
  tar: --exclude ‘internal’ has no effect
  tar: --exclude ‘passwordfile.txt’ has no effect
  tar: Exiting with failure status due to previous errors
  

• Others said to put it before the other tar switches, but that only puked:

  tar --exclude="internal" --exclude="passwordfile" cfvz ../backup.tar.gz *
  tar: You must specify one of the '-Acdtrux', '--delete' or '--test-label' options
  Try 'tar --help' or 'tar --usage' for more information.
  

• Still others said to put it after the other switches, but before the archive name:

  tar cfvz --exclude="internal" --exclude="passwordfile" ../backup.tar.gz. *
  tar: Removing leading `../' from member names
  tar: ../backup.tar.gz.: Cannot stat: No such file or directory
  contact.html
  db.bkup.tar.gz
  index.html
  internal/
  internal/import.db
  internal/scratchfile.txt
  passwordfile.txt
  public/
  tar: Exiting with failure status due to previous errors
  

  This was the worst so far, as it didn't work, and now I've got a --exclude=internal file in the directory!

  ls
   contact.html   db.bkup.tar.gz  '--exclude=internal'   index.html   internal   passwordfile.txt   public
  

  To remove a file that starts with a dash -, you need to use rm -- or rm ./ before the filename.

  rm -- --exclude=internal
  

• After much trial and error, I figured out that the --exclude needs to go in between the archive name and what is being added to the archive:

  tar cfvz ../backup.tar.gz --exclude="internal" --exclude="passwordfile.txt" *
  contact.html
  db.bkup.tar.gz
  index.html
  public/
  

References

GNU tar: an archiver tool - 6.4 Excluding Some Files https://www.gnu.org/software/tar/manual/html_node/exclude.html