[openstack-dev] [TripleO] Forming our plans around Ansible

Giulio Fidente gfidente at redhat.com
Wed Jul 12 06:04:49 UTC 2017

On 07/12/2017 01:53 AM, James Slagle wrote:
> On Tue, Jul 11, 2017 at 5:53 PM, Steve Baker <sbaker at redhat.com> wrote:


>> I think its important that we allow full support for both mistral-driven and
>> manually running playbooks. If there was no option to run ansible-playbook
>> directly then operators would miss one of the main benefits of using ansible
>> in the first place (which is leveraging their knowledge of inventory,
>> playbooks and roles to deploy things).
> +1, I like this idea as well. If you have a few minutes could you
> summarize it here:
> https://etherpad.openstack.org/p/tripleo-ptg-queens-ansible

note that this is how option (3) currently operates; it runs an 
unmodified version of ceph-ansible, installed on the undercloud so what 
the user needs to do on failure is to look for the mistral task that 
triggered the playbook and rerun the command

what it misses, as pointed by Steven, is a dump of the execution 
environment, that provides the extra_vars given to the playbook ... heat 
has this data, it should be possible to dump it in a file on the 
undercloud if we want to

I believe Steven is, with (4), trying to improve/reuse the mechanim

> I'm attempting to capture some of the common requirements from this
> thread for discussion at the ptg so we can consider them when choosing
> solution(s).

>> What would be nice is when a heat->mistral->ansible upgrade step fails, the
>> operator is given an ansible-playbook command to run which skips directly to
>> the failing step. This would dramatically reduce the debug cycle and also
>> make it possible for the operator to automate any required fixes over every
>> host in a role. This would likely mean rendering out ansible config files,
>> playbooks, (and roles?) to the operator's working directory. What happens to
>> these rendered files after deployment is an open question. Delete them?
>> Encourage the operator to track them in source control?

interesting question, as long as we run playbooks from a filesystem, I 
suppose users can make customizations without "changing" anything in 
tripleo ... this is how we tested some of the ceph-ansible fixes!

for upgrades we should maintain the tasks outside the templates do be 
able to do that though, assuming we want users to customize the upgrade 
Giulio Fidente

More information about the OpenStack-dev mailing list