Ansible lineinfile duplicates line

Question

I have a simple file at /etc/foo.txt. The file contains the following:

#bar

I have the following ansible playbook task to uncomment the line above:

- name: test lineinfile
  lineinfile: backup=yes state=present dest=/etc/foo.txt
              regexp='^#bar'
              line='bar'

When I first run ansible-playbook, the line gets uncommented and the /etc/foo.txt now contains the following:

bar

However, if I run ansible-playbook again, I get the following:

bar
bar

If I run it yet again, then the /etc/foo.txt file will look like this:

bar
bar
bar

How to avoid this duplications of lines? I just want to uncomment the '#bar' and be done with it.

Answer 1

You need to add backrefs=yes if you don't want to change your regular expression.

- name: test lineinfile
  lineinfile: backup=yes state=present dest=/etc/foo.txt
              regexp='^#bar' backrefs=yes
              line='bar'

This changes the behavior of lineinfile from:

 find
 if found
   replace line found
 else
   add line

to:

 find
 if found
   replace line found

In other words, this makes operation idempotent.

Answer 2

The problem is the task's regex only matches the commented out line, #bar. To be idempotent, the lineinfile task needs to match both the commented and uncommented state of the line. This way it will uncomment #bar but will pass bar unchanged.

This task should do what you want:

- name: test lineinfile
  lineinfile: 
    backup=yes
    state=present
    dest=/etc/foo.txt
    regexp='^#?bar'
    line='bar'

Note the only change was adding a "?" to the regex.

Answer 3

See https://github.com/ansible/ansible/issues/4531.

The solution is to not replace the commented out line, but to add an additional line, while keeping the original there.