David Ash wrote:
> Regarding the release cycle talked about at the UG last night....
> Can this be campared to the server release cycle of linux distributions?
> Can't find the exact values of release cycle and duration of support but...
> eg. Redhat has it's server class distributions "redhat enterprise ES/AS"
> with a release cycle of years and supports for something like 4 years or
> so and it's hobbyist distribution "fedora" with a 3-9month release cycle.
> When redhat did this, many clients complained as many hobbyists wanted
> secure public servers and didn't want to upgrade the OS every 3-9months,
> they don't mind doing bug fix/security updates but a full OS upgrade
> was a problem for a server class linux distrubution.
A MySQL Network subscription would get those users access to binaries on
a slower release cycle with only serious/security bugs fixed.
The key objective of the community builds is "release early and often"
to get the latest and greatest. If you run a production system on it,
you know what you're doing and you can make decisions on when to
upgrade, to what version, and how.
What you describe above is essentially users (which you call hobbyists)
that are running servers that behave like "enterprise servers" but they
don't want to take the consequences and either put in their own admin
effort, or subscribe to RH or MySQL services designed for that purpose.
I wouldn't call them hobbyists, given their work environments.
And anyway, they can't have it both ways... you either spend time to
save money, or spend money to save time. Free as in freedom.
> So will distributions just release mysql major updates more frequently?
Most distros are fairly conservative in moving to a new major RDBMS
release. I don't see such a move happening as an "update" within a
release cycle of a distro.
Debian for instance tends to backport security fixes into older
branches, regardless of what the upstream vendor does.
> Or will distributions start to support other open source databases
> instead of mysql?
Guh? Most distros have most if not all OSS dbs included.
I don't see why any of the above would make a difference to that.
Distros like the increased predictability of the MySQL releases, and the
improved communications with people on our end. This allows them to plan
stuff better, follow-up on problems, and allow time for enough testing.
Users who want access to stuff quicker than the release cycle of their
distro, grab their stuff from mysql.com. That's always been an option.
So to summarise, I don't think it's exactly the same as with Linux
distros. The issues of a distro are different to that of infrastructure
components such as MySQL.
Arjen Lentz, Community Relations Manager, MySQL AB
Program Chair, MySQL Users Conference
MySQL Users Conference 2006 (Santa Clara CA, 24-27 April)