[elrepo] LTS kernel EOL

[Phil Perry]

On 23/10/17 19:49, Shawn Asmussen wrote:
> Well, for me the compelling reason would be the same reason to use an 
> LTS kernel over the stock RHEL/CentOS one. The ability to use a less out 
> of date kernel. I think that if enough years pass, people who are 
> wanting to switch away from the kernel that ships with the OS will 
> become less interested in switching to an LTS kernel that is itself 
> several years old. Now, we can always take the most recent source 
> package available for the branch that we want and modify it to compile 
> our own RPMs for the current version of that branch (Like you can 
> currently do if you want a current 4.9 kernel), so it's not like we 
> don't have any options to work around this, but I just thought that an 
> increasingly aging LTS kernel version would begin to become less useful 
> over time.
> 

What you are describing doesn't really sound like you need a Long Term 
Support (LTS) kernel, it sounds like you really just want a newer kernel 
so perhaps kernel-ml is a better fit to your needs?

Paradoxically, our kernel-lt package can actually end up providing an 
'older' kernel experience than the stock distro kernel. For example, 
kernel-lt is rigidly 4.4 whereas much of the distro kernel is now newer 
than 4.4 (e.g, wireless stack is from kernel-4.11) so what was once 
'newer' is no longer necessarily so.

But as with the distro Enterprise Linux kernel, it's not about older or 
newer, it's about providing Long Term Support and a platform others can 
develop around is a stable and dependable way, knowing it's not going to 
change. And that to me is a far more compelling reason not to change it.