[PLUG] RH 7.x EOL

Terry Griffin griffint at pobox.com
Wed Apr 2 20:32:02 UTC 2003


On Wednesday 02 April 2003 07:27 pm, Rich Shepard wrote:
> On Wed, 2 Apr 2003, Terry Griffin wrote:
> > Well, thanks to all of you for pointing this out, but it doesn't help me
> > one bit as my boxes are mostly 7.0.
>
> Terry,
>
>   This lets me raise the concern I had when I read the e-mail this morning.
>
>   Based on Red Hat's lumping 7.0 with 6.2 suggests to me that there were
> _major_ changes from 7.0 to 7.1. And, that the releases from 7.1 forward
> are similar enough to be supportable through the end of this year.
>
>   So, my question is whether an upgrade of all your boxes to 7.1, .2, or .3
> would be equally as painful as an upgrade to 8.0 or 9.0. And, if the
> changes between 7.0 and 7.1 were so significant, why didn't they warrant a
> major number change rather than a minor number change?
>

Well, this isn't the discussion I wanted to have, but here's the history
as I understand it.

7.0 was going to be the first system with a 2.4 kernel, but 2.4 took a lot 
longer than expected, so 7.0 had a 2.2 kernel and 7.1 was the first to
have a 2.4 kernel. That's the split between 7.0 and 7.1.

The reason that 7.1 was not called 8.0 was because it was binary compatible
at the application layer. The kernel version generally does not impact binary
compatibility, except for really big things like the transition from a.out to
ELF.

Red Hat's pattern until now was to roll the major number when there was
something that broke binary compatibility, usually something very basic like
a new version of glibc, gcc, or both.

As for upgrading my boxes. It would just be a big pain in the backside
irrespective of the size of the leap, and for no good reason other than to
follow with along with Red Hat's grand plan for making money from bug
fixes. These boxes are all doing their respective jobs perfectly well.
Taking them out of service for upgrades makes no sense if I can avoid it.

Terry





More information about the PLUG mailing list