F22 Self Contained Change: Database Server Role
Jaroslav Reznik
jreznik at redhat.com
Tue Jan 20 15:23:03 UTC 2015
----- Original Message -----
> Once upon a time, Stephen Gallagher <sgallagh at redhat.com> said:
> > One of the core focuses of Fedora Server is to simplify things. It's
> > meant to help less-experienced users of Linux get up and running with
> > common activities more quickly. Providing the "PostgreSQL" role and the
> > "MariaDB" role means that we've forced the user to do additional
> > research to figure out what they want. However, if we name one "Database
> > Server", we are implicitly telling the user: "use this one, unless you
> > have a specific need".
>
> Well, but a user will still have to do that research. A database isn't
> like a browser or word processor; it doesn't exist in a vacuum and one
> database engine can't just replace another at will. Some programs
> support a wide variety of database engines, and some don't. For
> example, for good or bad, many PHP-based things assume MySQL; some can
> be configured otherwise, but most default to MySQL (and that may be all
> the developers actually test).
Maybe this could be a very nice LAMP role. Or in our case FAMP :).
> Database engines are probably one of the least interchangeable pieces,
> so choosing _any_ (I'd say the same thing if this was a proposal to use
> MariaDB) as "THE database engine" is poor IM(very)HO. It isn't about
> promoting one engine over another, it is just that none are really the
> one engine to rule them all. Given that, I don't see a reason to
> declare any engine as the one true Database Server. I think any role
> for a database should have the engine name in the Role.
I understand what you mean but also - if someone is going to deploy
application that depends on specific DB, it would probably mean own
installation based on app requirements.
Jaroslav
More information about the devel
mailing list