[CentOS] How do we handle panics? Bug it here or RH or ignore it (been around *long* time apparently)?

Mon Sep 4 14:29:28 UTC 2006
William L. Maltby <CentOS4Bill at triad.rr.com>

On Sat, 2006-09-02 at 14:31 -0400, William L. Maltby wrote:
> "Unable to handle kernel paging request". I've saved the OOPS data from
> the logs for 6 panics since the 4.4 update.

s/6/10/  # Now

> <snip>

> I've started reading what I can find on crash to see if I can get the
> whole thing. Do we have a kernel with debug symbols and associated?

Haven't stayed running long enough to pursue this (did take time for
real life stuff, seems more manageable)

> 
> I don't know if I'll pursue this. Disabled swap and that seems to be
> working. I'm fortunate I can do this. Others may not be so lucky. I'd
> like to turn swap back on, but I'll probably have to disable all that
> read-ahead stuff again and find the thread on the swappiness switch and
> try and find the list of kernel parameters *again* (I know there in here
> somewhere)...

Have tried several combinations without success. Swapoff, 0
> /proc/sys/vm/swappiness, those two without each other, disabled the
readahead and readahead_early stuff, running in run level 3 only,
nothing's worked to keep it up. Need Robo-Viagra here.

Am currently running with swappiness at 10, swap enabled both readaheads
disabled. This is all the stuff that was gleaned from the prior CentOS
list discussions.

Anyone got any other things I might try? I know a fix is not yet
available (if my googling was extensive enough and I missed nothing) but
I would like a work-around that might keep it up more than 24 hours
about 90% of the time.

One new piece of info: a lot of the OOPS, but not all, have started
after the machine was idle and I touch the keyboard to bring the things
back to life (AFAIK, just a screen-saver, blanked, going). So I turned
off the BIOS ACPI stuff. Since I know zilch about the ACPI stuff, I've
begun reading /usr/share/doc stuff (kernel params and pm) to see what I
might disable in there. Maybe that will help some.

But if anyone has a couple suggestions regarding that, I might get the
docs read faster (fewer boots). I would appreciate it.

> Enough griping for today. Do I bugzilla CentOS, RH or ignore it?

Since I'm somewhat new here, and this is a recognized problem in the
community (if my googling is correct), I am still uncertain how I should
deal with this, other than the workaround. Can someone please respond to
my simple question: "Do I bugzilla CentOS, RH or ignore it?"

<snip>

--
Bill