[Nottingham] Various kernel oops

[Graeme Fowler]

On Tue, 2008-05-27 at 23:27 +0100, Martin wrote:
> OK, so these are my first kernel oops...

Either you've been lucky or you have $deity inside your boxes :)

> Anyone seen anything like this and any good (or quick!) fixes?

Not specific to your hardware, no, but I've seen enough "unable to
handle kernel paging request" to have a general feel for it: in my
experience, your system is driving itself into swap and experiencing a
problem. It could be that there simply isn't enough (although that
normally triggers the kernel's "OOM killer" first) or that the device(s)
the swap lives on just aren't fast enough.

Having swap on a very busy RAID5 array, for example, where swap is a
slice of the same logical drive as /var and /var is being hammered by
mail, web requests, cgi/php lock files, MySQL and logging all at the
same time can result in a paging request to swap simply taking too long
to fulfil.

If you turn off swap, then the symptoms go away (as you notice) but can
recur (as you also notice) when physical RAM gets full.

As an aside, are you using a processor with hyperthreading? If you are,
turn it off and see if that makes a difference. Also don't overclock.

Graeme