Service option must_exist disappeared

[Hubbitus]

ISSUE TYPE

• Bug Report

ANSIBLE VERSION

2.0.0-0.6.rc1

COMPONENT NAME

service_module

It was mentioned as became available in ansible 2.0 in doc: http://docs.ansible.com/ansible/service_module.html . It also was work as described.

In recent versions (2.0.0-0.6.rc1) it produce error:

fatal: [egais.db.slave1]: FAILED! => {"changed": false, "failed": true, "msg": "unsupported parameter for module: must_exist"}

And doc does not contain it.

Is that dropped? Why?

[bcoca]

I did this, need to look at my notes for exactly why, putting it in my queue to check once back home to review my notes

[peterhoeg]

There is a PR here ansible/ansible-modules-core#2816 which solves it. I'm manually applying the patch here.

[dailyherold]

@bcoca found this issue after following up on some old TODO's I had in my repo regarding must_exist. Any insight into what happened? Your time is much appreciated!

[yitzikc]

I still have this issue as of 2.0.2, requiring an ugly workaround. Any suggestions as to how it can be expedited?

[bcoca]

So this was a feature we attemtped to add in devel, but never released, I had to remove it as it seemed to break several existing plays and I never had time to look at the corner cases.

In any case there is easy existing workaround using failed_when:

[phemmer]

I had a related ticket (#16134) that was closed (objectionably) as a duplicate of this one. So I'm copying the core for that ticket here:

The issue is that ansible throws an error when stopping/disabling an unknown service. The must_exist option would be able to fix this use case, but the behavior should be that it is automatically enabled, though only when stopping or disabling the service. Thereby restoring the behavior of ansible 1.9.4
Satisfaction of this requirement does not require a must_exist option. It could be resolved without implementing must_exist as the behavior would be automatic (hence why it is not a duplicate).

[ansibot]

@Hubbitus Greetings! Thanks for taking the time to open this issue. In order for the community to handle your issue effectively, we need a bit more information.

Here are the items we could not find in your description:

• issue type
• ansible version
• component name

Please set the description of this issue with this template:
https://raw.githubusercontent.com/ansible/ansible/devel/.github/ISSUE_TEMPLATE.md

click here for bot help

[Hubbitus]

I had fill ansible version, issue type and component name.

[Gunni]

this bug still active? the PR was pulled and i could not see an alternative, right now all disables of non-existant throw errors which i have to ignore

[sivel]

We have recently made the decision that we plan on leaving the behavior as is (failing when the service does not exist)

As discussed in #35293 the correct option here is to use either ignore_errors: true or a failed_when statement to prevent an error.

Logs of that meeting:
https://meetbot.fedoraproject.org/ansible-meeting/2018-01-25/ansible_core.2018-01-25-15.00.log.html

If you have further questions please stop by IRC or the mailing list:

• IRC: #ansible on irc.freenode.net
• mailing list: https://groups.google.com/forum/#!forum/ansible-project

[Gunni]

So, you are saying that check mode should always say changed and non-check should always error for this.

That just feels very unpolished..

[Gunni]

I solved my case this way, this feels dumb but i'm limited by what you give me... And it's not portable to systems without systemd.

p.s. before you ask, i'm not running the machines naked, i already enabled nftables when this runs.

- name: Check if services exist
  command: "systemctl status {{item}}"
  check_mode: false
  changed_when: false
  failed_when: killable_services.rc not in [0, 3, 4]
  with_items:
  - firewalld
  - ip6tables
  - iptables
  register: killable_services

- name: Stop and disable services
  systemd:
    name: "{{item.item}}"
    state: stopped
    enabled: false
  with_items:
  - "{{killable_services.results}}"
  loop_control:
    label: "{{item.item}}"
  when:
  - item.rc != 4

[bcoca]

much simpler:

- service:
  ...
  register: myres
  failed_when: myres is failed and not 'Could not find the requested service' in myres.msg