Error checking for daemon/service restarts

[David Marshall]

Hi, 

I want to make sure I am going about this in the cleanest and easiest way possible. Maybe someone else has done this before or has some better ideas for me :)?

When a service is restarted i.e. "service networking restart", I want to check for stderr and stdout to make sure there were no issues with the ansible push.

I am using "nested" handlers to accomplish this atm:

------------------------------

Here is the Playbook/Handlers:

---

- hosts: spine1

  user: root

  gather_facts: false

  tasks:

    - name: copy interfaces file

      copy: src=interfaces dest=/etc/network/interfaces

      notify: reload networking

    - name: copy quagga file

      copy: src=quagga dest=/etc/quagga/Quagga.conf

      notify: reload quagga

  handlers:

    - name: reload networking

      shell: "service networking reload"

      register: networking_reload

      notify: print networking reload error

      changed_when: "networking_reload.stderr != ''"

    - name: reload quagga

      shell: "service quagga reload"

      register: quagga_reload

      notify: print quagga reload error

      changed_when: "quagga_reload.stderr != ''"

    - name: print networking reload error

      debug: msg="{{networking_reload.stderr}}"

      when: networking_reload.stderr is defined

      failed_when: "networking_reload.stderr is defined"

    - name: print quagga reload error

      debug: msg="{{quagga_reload.stderr}}"

      when: quagga_reload.stderr is defined

      failed_when: "quagga_reload.stderr is defined"

---------------

Tested scenerios:

config file is different with error:

root@wbench:~/consulting/testing# ansible-playbook test.yml --diff

PLAY [spine1] *****************************************************************

TASK: [copy interfaces file] **************************************************

changed: [spine1]

TASK: [copy quagga file] ******************************************************

ok: [spine1]

NOTIFIED: [reload networking] *************************************************

changed: [spine1]

NOTIFIED: [print networking reload error] *************************************

failed: [spine1] => {"failed": true, "failed_when_result": true, "verbose_always": true}

msg: error: /etc/network/interfaces: iface swp52: unsupported keyword (this_should_not_be_here,_it_makes_no_sense._GAH)

FATAL: all hosts have already failed -- aborting

config file is different with no error

root@wbench:~/consulting/testing# ansible-playbook test.yml --diff

PLAY [spine1] *****************************************************************

TASK: [copy interfaces file] **************************************************

changed: [spine1]

TASK: [copy quagga file] ******************************************************

ok: [spine1]

NOTIFIED: [reload networking] *************************************************

ok: [spine1]

config file is the same with error:

(no change to handler isn't called) (only way around this is to ifreload on every run)

config file is the same with no error:

root@wbench:~/consulting/testing# ansible-playbook test.yml --diff

PLAY [spine1] *****************************************************************

TASK: [copy interfaces file] **************************************************

ok: [spine1]

TASK: [copy quagga file] ******************************************************

ok: [spine1]

------------------------------------------------------------------------------------------------------------------------------------

It looks like the service module does not actually return stdout/stderr.

Here is a run using the service module and just printing the returned var:

ok: [spine1] => {

    "var": {

        "networking_reload": {

            "changed": true,

            "invocation": {

                "module_args": "name=networking state=reloaded",

                "module_name": "service"

            },

            "name": "networking",

            "state": "started"

        }

    }

}