[rdo-list] Packstack refactor and future ideas

[Alan Pevec]

> What is the difference between creating a new "unsupported" additional
> tool and marking multihost capabilities in packstack as
> "unsuported/untested"?.

Looks like one of my reply didn't reach rdo-list (I got
<rdo-list at redhat.com>... User unknown ??)
Difference is that no matter what deprecation warnings you put people
will keep using features until they're really gone and those features
are out of scope.

> Current code enabling multihost in packstack is working and i have

It still adds complexity vs assuming local execution i.e. now
Packstack does ssh even to localhost, right?

> doubts about the expected wins in time supporting multihost in
> packstack versus creating a new ansible tooling and modifying
> packstack to work fine in this new "externally orchestrated mode".

Win is the clear focus for the tool and local execution without
messing with ssh.
I like Javier's proposal that would keep controller, network, compute
then you just need to execute packstack on multiple machines with the
same answer file, manually or using Ansible.
In that case, "new ansible tooling" is really just an example playbook in docs/.

Cheers,
Alan