Using serial strategy for playbook imported in a toplevel playbook causes entire play termination if some host is unreachable

[Andrey Agenosov]

Hello all.
Here is a simplified example:

- import_playbook: test_unavailable_node_different_strategy.yml

- hosts: 127.0.0.1
  connection: local

# The question is would we end up here in case of imported playbook (above) touches unavailable host
  tasks:
    - debug: msg="Playbook finished"Введите код...

And the code for imported playbook:

- hosts: all
  # Comment the line below and everything would be fine
  serial: 1

  tasks:
    - debug: msg="{{ inventory_hostname }}"Введите код...

So when I'm launching something like this

ansible-playbook -i '127.0.0.1,10.0.0.1' use_unavailable_nodes.yml --ask-pass -u andreyВведите код...

when the 10.0.0.1 host is unreachable, the entire play aborts before reaching the task to be performed locally (see main playbook).

If you comment 'serial: 1' then the local action in the main playbook also would be performed.
Why there is such behaviour? Is this a bug or I'm missing something?

[Marcos Alano]

I ran a playbook on batches and one host failed but just that batch stopped.

> --
> You received this message because you are subscribed to the Google Groups
> "Ansible Project" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to ansible-proje...@googlegroups.com.
> To post to this group, send email to ansible...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ansible-project/3c31fa30-7b75-4fbc-9e3b-ece572980f06%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
Marcos H. Alano
Linux System Administrator
marcos...@gmail.com