canonical-ci-engineering team mailing list archive

[Evan Dandrea <evan.dandrea@xxxxxxxxxxxxx>]

FYI


---------- Forwarded message ----------
From: Tom Haddon <tom.haddon@xxxxxxxxxxxxx>
Date: 11 November 2013 09:45
Subject: Re: Charms, charm-helpers, and salt vs ansible
To: Chris Johnston <chris.johnston@xxxxxxxxxxxxx>
Cc: Evan Dandrea <evan.dandrea@xxxxxxxxxxxxx>, Michael Barnett
<michael.barnett@xxxxxxxxxxxxx>, Andrew Glen-Young
<andrew.glen-young@xxxxxxxxxxxxx>


On 08/11/13 18:44, Chris Johnston wrote:
> Hey Tom.. I was told that you would be the person to ask. I know that
> webops wants charms to use charm-helpers, however I'm not sure if there
> is a particular direction that is preferred when it comes to Salt vs
> Ansible. Is there a preference, or any sort of reason to use one over
> the other?

Hi Chris,

I'm not currently heading up the WebOps team - that's Michael Barnett,
although due to team rotations it'll be Andrew Glen-Young in a few
weeks, so I've cc-ed them both here.

We don't have an absolute policy saying "you must use charm-helpers" but
we do strongly advise it for python-based charms (which we strongly
prefer for any non-trivial charm). In terms of Salt vs. Ansible we
haven't used or seen charms using either of those yet, but I don't
believe we'd have a problem with either of them, assuming they're well
implemented and don't contradict
https://wiki.canonical.com/InformationInfrastructure/IS/Policies/Prodstack.

Thanks, Tom

> Thanks,
>
> cJ
>
> --
> Chris Johnston <chris.johnston@xxxxxxxxxxxxx
> <mailto:chris.johnston@xxxxxxxxxxxxx>>
> Software Engineer - CI Engineering Team
> Canonical Ltd.
> www.ubuntu.com <http://www.ubuntu.com>