Ansible replace or lineinfile?

Question

All,

I have been trying to figure something out all day, i got ot working with the shell module and sed but cant do it with a more native way such as replace or lineinfile. This is what i am trying to do:

I have a file that has a few lines and some lines start with a different string but others with the same which are the ones i want to edit, then all the lines at some point have a particular string and i want to edit this string only on the lines that start with that particular string.

Here is a more visual example:

The file by the way is /etc/fstab and looks something like this

Aaaaa:/mount1  /mountpoint1 nfs mount-options 0 0
Aaaaa:/mount2  /mountpoint4 nfs mount-options 0 0
Bbbbbn:/mount1  /mountpoin9  nfs mount-options 0 0

As you can see in this example the mount options which is what I want to change are identical however I want to edit only the ones in lines that start with aaaa for example.

Any clues how to do this with a module like replace or lineinfile and not sed using the shell?

Thanks

UPDATE: Let me clarify that the stuff after aaaaa: until nfs mount-options 0 0 could be anything, as the mount folders and mount points on the server could be anything. Also this is how I got it to work using the shell module, still looking for a more native module instead of shell:

  - name: Inserting new mounting options
    shell: sed -i '/^aaaaa/ s/mount-options/new-mount-options/g' /etc/fstab

In the above example I successfully target lines starting with aaaaa and replace the mount-options to different ones while keeping the 0 0 at this end.

I am not trying to mount anything, I am trying to edit a line particular line in the file, lets forget that the file is /etc/fstab, i just want to edit those lines that start with a certain word in them and not the others. — Yohn Gutierrez, Aug 13 '19 at 22:42
Can you also add the working shell module in your question? I am a bit confused about what you are trying to achieve — Matt P, Aug 13 '19 at 23:37

score 0 · Answer 1 · answered Aug 14 '19 at 10:21

0

Replace fits better the use-case. See the example below. For example, let's

set dump fsck to 1 1 where lines begin with Aaaaa
set fstype to ext4 where lines begin with Bbbbbn

The task below

- replace:
    path: /scratch/fstab
    regexp: "{{ item.regexp }}"
    replace: "{{ item.replace }}"
  loop:
    - regexp: '^Aaaaa(\S+)\s+(\S+)\s+(\S+)\s+(\S+)\s+(\S+)\s+(\S+)(.*)$'
      replace: 'Aaaaa\1 \2 \3 \4 1 1'
    - regexp: '^Bbbbbn(\S+)\s+(\S+)\s+(\S+)\s+(\S+)\s+(\S+)\s+(\S+)(.*)$'
      replace: 'Bbbbbn\1 \2 ext4 \4 \5 \6'

gives

$ cat fstab
Aaaaa:/mount1 /mountpoint1 nfs mount-options 1 1
Aaaaa:/mount2 /mountpoint4 nfs mount-options 1 1
Bbbbbn:/mount1 /mountpoin9 ext4 mount-options 0 0

^{The regex are cumbersome. I can't get working Capturing repeating subpatterns in Python regex.}

answered Aug 14 '19 at 10:21

Vladimir Botka

58,131
4
32
63

I am a little lost at your regex there, i will try to take it apart and see if it can work for me – Yohn Gutierrez Aug 14 '19 at 12:31
The regex comprises 7 groups; 6x (\S+) non-whitespace chars, and 1x (.*) any chars separated by at least one whitespace \s+. These 6 groups correspond to the fields in fstab. It can be easily changed. Put any regex instead of a group. For example to select fstype `nfs` use '^(\S+)\s+(\S+)\s+(nfs)\s+(\S+)\s+(\S+)\s+(\S+)(.*)$' – Vladimir Botka Aug 14 '19 at 12:46
Ah this might help with what i am trying to do becase i just found that the mount options are different everytime so i need to target groups instead. I will give it a go. – Yohn Gutierrez Aug 14 '19 at 13:14

Ansible replace or lineinfile?

1 Answers1