Page faults every few days
kstewart at owt.com
Sun Dec 14 01:42:03 PST 2003
On Sunday 14 December 2003 01:16 am, jaime at snowmoon.com wrote:
> On Sun, 14 Dec 2003, Kent Stewart wrote:
> > Well, there is this piece of info in /usr/src/UPDATING
> > 20030904:
> > Between August 9th and August 30th, a bug existed in the i386
> > virtual memory system which could cause panics under load.
> > Anyone running a kernel built between those dates is advised
> > to update at the earliest possible convenience.
> How'd I miss that?! OK, this is very useful and I really
> appreciate it. Its just that I was sure that:
> 1) The panics started before then
> 2) I read UPDATING
> I'm obviously wrong on #2, so now I'm less certain about #1. I'll
> get on this a.s.a.p. and see what it does.
You could have been seeing them; however, I think a point you should remember
is that when you are having problems, you have to get to an environment that
someone can reproduce to duplicate your problem. You probably can't do that
on a prerelease. I know I couldn't be talked into going back with the system
I do tests on. It is also much more likely that some one is running something
like RELENG_4_9 or a current -stable. They really don't have to do anything
other than simulate your load and see if the system panics.
I would have bet anything from a flaky power supply to cooling fans until I
saw the notice in UPDATING. I had several machines running kernels created
during that period and none of them paniced but IIRC, there were a number of
them panicing at the time. I place more creadence on your failure than my
More information about the freebsd-questions