At 10:07 AM -0500 8/19/06, Les Mikesell wrote: >On Sat, 2006-08-19 at 06:44, Matthew Miller wrote: > >> > > > Centos 3.x is still a good server choice. I've always considered >> > > > linux kernels to be stable around version x.x.20 or so. >> > > It's different now. >> > Do you mean that with no odd-numbered development branch the >> > 2.6.x line may never stabilize? >> >> Even and odd no longer have anything to do with it. The 2.6.x.y releases get >> a bit of stabilization, but mostly it's full steam ahead. The idea is that a >> litte bit of change every release is better than a gigantic lump of changes >> with a 2.8.0 which then requires 'til 2.8.20 to run smoothly. > >That's good for the developers but bad for people who don't >like surprises when applying needed distribution updates >to their servers - which is why most of mine are still running >CentOS 3.x. With the odd/even cut, Linus was always way too >optimistic about declaring the version to be stable, but >by about x.x.20 it got there. With little changes introduced >all the time and no one backporting security/bugfixes into >a well-tested release, how can anyone pick and maintain a >production version? I read something on LKML, I think, that bug fixes were being backported now. So even after 2.6.x+1.0 comes out, there may still be 2.6.x.n releases with bug fixes. I suppose that Fedora stays with the current x as it goes up, but there may be bug fixed vanilla kernels with older x. ____________________________________________________________________ TonyN.:' <mailto:tonynelson@xxxxxxxxxxxxxxxxx> ' <http://www.georgeanelson.com/>
