maas-devel team mailing list archive

[Andreas Hasenack <andreas@xxxxxxxxxxxxx>]

On Wed, Feb 5, 2014 at 2:15 PM, Andres Rodriguez <
andres.rodriguez@xxxxxxxxxxxxx> wrote:

> I don't have any strong opinions on what the possible fix might be, but
> I'm wondering if an appropriate fix for this should better be made on the
> actual postgresql charm? I mean, we have other services that do not have
> these type of issues, so it seems to me that the appropriate place to have
> a fix for this, is on the charm side.
>
>
Look at the branch attached to the bug, it's against the postgresql charm
and it checks to see if the unit name is a cname or not. We couldn't have
postgresql not deploying for months :)

I consider it a workaround. CNAMEs are for humans, it's the www in
www.canonical.com. And the fact that ssh does the same check tells me it's
the right thing to do when it comes to access controls.