On 29/07/2011 20:19, Herbert Poetzl wrote:
> here the status so far:
> all the money so far is on 2.6.32.x while the majority
> votes for 3.0.x ...
Put me down for $500 on 3.x and newer (I'm on holiday in the sticks -
will send money shortly). As Adrian Reyer says, kernel development
model is basically that when you want to buy some new server hardware or
use some new iptables feature, you tend to be required to go with the
latest, greatest kernel...
Pax are arguably a similar model to vserver. Their solution is to
follow mainline closely, bug fix for a while on a particular release,
but largely try and stick with the latest.
Look at the buzz which accompanied the RT patch suddenly supporting some
4+ year newer kernel... I think a project will look dead and perhaps
become dead if you *only* nail your supported versions in the past.
I also don't want to upgrade kernels every 10 minutes, so my solution is
to add the Pax patch to try and avoid needing to upgrade kernels so
regularly. Machines then tend to sit there without reboots for some
years... So it really does me no benefit for someone to carry on
developing for that old kernel version - I'm not going to notice...!!
Next time I need a new kernel either for a reboot or a new machine,
probably I will evaluate something much more recent anyway...
Those of you voting for 2.6.32 or .38 or whatever, try and notice the
different between what you are actually running on your servers today
and whether it benefits you in the slightest to keep developing for that
version... Surely your 2.6.32 machine is running fine right now? (Mine
are) Surely the next time you care to install a new kernel it will be
because you have upgraded to some new version of the distro that is
likely based on 3.x? Don't you prefer to see 3.x be the one developed
*now* so that it's ready for you to move to next year?
I feel there is a lot of backwards looking here. Think where you want
to be in 1-2 years time, not where you want to be last week
Cheers
Ed W
Received on Sat Jul 30 14:13:46 2011