reviving Fedora Legacy

Les Mikesell lesmikesell at gmail.com
Wed Oct 15 13:12:04 UTC 2008


David Woodhouse wrote:
> 
>>> I really don't see how a Fedora Legacy can be maintained.  If the 
>>> goal is increased stability and security patches, you need to 
>>> guarantee that you have folks supporting backpatches to the kernel, 
>>> glibc, firefox, evolution, openoffice, and several other large and 
>>> complex packages.  Incorporating new security patches into old 
>>> baselines is *hard*. Plus Fedora would "fork" a new release every 6 
>>> months.  How many legacy Fedora's would be retained?  At some point 
>>> it seems the legacy volunteer force would saturate and legacy 
>>> Fedora's would have to start dropping off every 6 months.
>> Why do we need to guarantee any more than active Fedora releases 
>> guarantee?  Forget backporting.  Just upgrade the package.  Take it 
>> from the current Fedora and rebuild it if necessary.
> 
> Once you start upgrading packages all over the place to a much newer
> version than was in the original release, you might as well just
> upgrade.
> 
> Seriously, I don't know why people are so scared of just _upgrading_, if
> new packages are acceptable.
> 
> I upgrade remote, headless machines with yum, and reboot them into the
> new distribution. Quite frequently. And I laugh at the people who say it
> doesn't work. It's a fairly fundamental part of my server management
> technique -- yes, I run Fedora on my servers.

What do you do when the upgrade kernel won't boot?   This sometimes 
happens even on updates within a version.

> Perhaps a better approach to this whole thing would be to educate people
> a little better that upgrades _do_ work, and they're generally fairly
> seamless. And to fix the occasional cases where they're not.

And supply them with spare test machines like you must use?

-- 
   Les Mikesell
    lesmikesell at gmail.com




More information about the fedora-devel-list mailing list