[plug] Kernel logs

[Craig Ringer]

>>One alternative if you care about your kernel logging in times of
>>despair
> 
> It seems that dealing with kernel-level fault scenarios is an area
> where other operating systems currently have an edge.

Quite likely so.

> I'm probably being parochial, but I don't like the idea of attempting to
> transfer cores over a network link as the first priority.

As summarised on kerneltraffic.org, there has been some argument on this 
among the kernel folks. They essentially settled on network crash dumps 
as the "safer" option, because of the risk of writing to the disk when 
the kernel is known to be in a bad state. I think there are patches out 
there to do both network and swap crash dumps, though.

> This sort of issue is obviously context-dependent, as is a somewhat
> related issue: most people probably think it sounds ideal that a reboot
> would give you a "fresh slate", but it is actually quite nice to have a
> system that does a more mild sort of reboot in the event of failure.

Sometimes, it can be. It also opens you up to issues though. If you're 
going to have a system like that you also need the ability to force 
resume without recovering state data from the last session, in case 
something in the last session broke the system. I'm not so sure this is 
a good idea in general purpose operating systems due to their complexity 
and relatively low quality. Windows likes to save aspects of the system 
state as you go, at least more than I really expect it to, and I've had 
that break things enough that I'm cautious about the whole idea.

Craig Ringer


_______________________________________________
plug mailing list
plug at plug.linux.org.au
http://mail.plug.linux.org.au/cgi-bin/mailman/listinfo/plug