Clear host errors not working as supposed - cannot retry task on unreachable host

[Listing]

HI al,

I am running a playbook on a set of hosts of which I expect them to be offline. I want to check every 10 seconds whether they are online or not and then output a message when all is running, or another message when all is not running after 10 retries for instance.

I am getting 'host unreachable' after which the unreachable host is being removed. I can continue to run other tasks on the hosts that were online (by using 'meta: clear_host_errors'), but I can't retry the task on the host that was not online.

Does anybody know how to keep trying (for instance 10 times) a task until all hosts have been successful? Is this possible?

[Kai Stian Olstad]

Have you check out wait_for?
https://docs.ansible.com/ansible/wait_for_module.html

and with "any_errors_fatal: true" you can at least stop the play if one
or more host doesn't come online.

--
Kai Stian Olstad

[Listing]

That works partially, the host is not being cleared but it keeps hanging on that unreachable host.

I managed to do a ping -c 3 to all the hosts in combination with retries=10. I only need to exclude certain host groups when I include all hostnames via: with_items: groups['all']

with_items: groups['all:!excluded_group'] does not seem to work. Have you got any idea how to exclude here? :)

[Kai Stian Olstad]

On 06. juli 2017 22:47, Listing wrote:
> with_items: groups['all:!excluded_group'] does not seem to work. Have you
> got any idea how to exclude here? :)

with_items: "{{ groups['all'] | difference(groups['excluded_group']) }}"

--
Kai Stian Olstad

[Listing]

Great, everything is working as I want it to be, thank you for your help!