On Mon, 8 Dec 2008, Scott Marlowe wrote:
> Are you familiar with this bug:
> https://bugs.launchpad.net/debian/+source/linux/+bug/245779
>
> It's the reason my latest db servers are running Centos 5.2, sadly.
> By the time I'd found the suggested workaround of setting a boot
> option of NO_HZ=y I was already migrated off ubuntu for db servers.
Don't want to drag Liraz's thread completely off-topic, thus the new
subject.
The response to that bug demonstrates one reason why I get a bit worked up
when people suggest using Ubuntu for any serious server work. Even when
bugs get fixed, it's far too often only via installing a newer kernel,
which puts you back to square one as far as testing goes. Ubuntu puts
minimal resources into back-porting kernel fixes into any earlier version,
LTS or not, because they're consumed with constantly churning out new
versions. The usual cut-and-paste response appears in your thread same as
it does in all the similar ones:
"The Ubuntu Kernel Team is planning to move to the 2.6.27 kernel for the
upcoming Intrepid Ibex 8.10 release. As a result, the kernel team would
appreciate it if you could please test this newer 2.6.27 Ubuntu kernel."
A good eye-opener if you don't believe who I'm characterizing things is
take a look at the location your bug ended up being parked at (and may
very well die at):
https://bugs.launchpad.net/~ubuntu-kernel-team/+assignedbugs
There you can gauge for yourself how concerned they are with fixing bugs
in older versions. You can't support yearly long-term support releases
and aggresively back-port fixes without way more resources dumped onto the
kernel team than Ubuntu has to apply. Even RedHat, who has a lot more
kernel engineers, doesn't even try. That's part of the reason why it took
more than two years between RHEL4 and 5. They were busy that whole time
backporting kernel fixes into the stable kernel, with major update drops
to it every six months, rather than just plowing ahead only worrying about
the newer ones.
I love Ubuntu on the desktop, but you combine its aggresive releases and
limited kernel fix backporting with how much general kernel testing
quality keeps going down and you get a grim combination. I've realized
this is just an unavoidable consequence of how much change the Linux
kernel is going under every single day. Nobody seem to care anymore about
focusing on any individual kernel version long enough to squash its bugs
right anymore; those will all get fixed in the next version, right?
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD