Re: [tlug] MySQL 4.1.15, or MySQL 5.0?

["Walter Hansen" <gandalf@example.com>]

> TLUG,
>
> I originally asked this question on the MySQL mailing list, but it got
> no response. Perhaps it was considered beneath them, as it's a pretty
> low level question. Perhaps the ever-helpful TLUG can set me straight:
>
> Currently I'm running MySQL 4.1.15 on my home computer where I do web
> development, and MySQL 4.1.13-beta on my hosting service. Actually, I do
> work on web sites that are hosted on many hosting services, but on only
> one of them do I have enough access to decide for myself if I want to
> upgrade the server.
>
> It seems that MediaWiki, something I'm currently experimenting with,
> requires MySQL 4.1.14 or above. So at the very least I need to upgrade
> MySQL on the hosting service.
>
> But I'm a little confused by the presence of MySQL 5.0. On the mysql.com
> web site, it promotes version 5.0 and barely makes any mention of
> previous versions. (Of course, in the support documentation there is a
> lot about previous versions, but I'm just talking about the site's
> promotional text.)
>
> Ordinarily, a new version of any software wouldn't be confusing. I would
> assume that whatever the latest version that the developers are making
> available is the one that is supported and stable and preferable to use.
>
> But none of my hosting services anything higher than 4.1.15. And my home
> computer, which runs Ubuntu, defaulted to installing 4.1.15 and the
> application update manager doesn't update it to 5.0.
>
> So if 5.0 is the "current" version of MySQL, why does it seem to me that
> it's not widely adopted?
>
> Would it be problematic now or later if I upgraded to 5.0?
>
> Would I have to soon upgrade to 5.0 if I upgraded to 4.15 now?
>
> Thank you for any advice.
>
> --
> Dave M G
>

Usually the major version changes have something more than a simple
update. Perhaps the actual table format on disk has changed. Anyway
upgrading from 4.x to 5.x is generally a big step where upgrading within
4.x is not a big deal.

Also these databases tend to hold a ton of data and it's generally very
lively. This makes it very difficult to do updates on a large operation.

The third reason is that companies like to wait and see how well the new
version will do before making the big jump. Distributions do the same
thing. I'm sure that 5.x is very well tested at this point and even had a
bunch of features that I really wanted at some point, but if the 4 series
is largely doing what you want, the push to move isn't great.

It's a little like php4 vs php5. PHP5 introduces OOP to PHP, but the
downside is that a lot of code written for 4 won't work on 5. This is such
a big split that many providers are putting both up and letting the
customers chose which to use by selective extensions. I imagine there are
similar worries about mysql 4/5. People don't want their million dollar
customer service apps to break when they upgrade.