How to loop over result and retry in the same task?

[Eric Chong]

For example,  if I want a task to pull ec2_instance_facts until any instance is in running state, and retry if none is running.

---

- hosts: localhost

  gather_facts: no

  connection: local

  tasks:

  - name: get ec2_instance_facts

    ec2_instance_facts:

    register: my_ec2

    until: item.state.name == 'running'

    with_items: '{{ my_ec2.instances }}'

But I am getting error, because my_ec2 is not defined yet.

TASK [get ec2_instance_facts] **********************************************************************************************

fatal: [localhost]: FAILED! => {"failed": true, "msg": "'my_ec2' is undefined"}

Is it possible to loop over the result and retry the task in a single task?  If I have to make a second task to loop over the result, I can't go back to retry the first task.  How does Ansible handle retry of multiple tasks?

Thanks

[Brian Coca]

you cannot loop over the results of your task that is generating those
results, you CAN loop over the results of one task in a second task.





--
----------
Brian Coca

[Eric Chong]

But I want to retry the first task until expected result.  I don't see a way for Ansible to retry a bock of tasks.

[Brian Coca]

you CAN retry the first task, use until condition, i dont see a block
of tasks ...





--
----------
Brian Coca