On Tue, 2010-09-21 at 16:02 +0200, Martin Sourada wrote:
On Tue, 2010-09-14 at 14:17 +0200, Martin Sourada wrote:
> On Mon, Sep 13, 2010 at 8:43 AM, Martin Sourada
> <martin.sourada(a)gmail.com> wrote:
> > On Mon, 2010-09-13 at 00:07 +0200, Martin Sourada wrote:
> >> On Sun, Sep 12, 2010 at 11:58 PM, Adam Williamson
> >> <awilliam(a)redhat.com> wrote:
> >>
> >> Check kernel logs and X logs after rebooting, see if there's
> >> crash
> >> notifications in either. (You'll want to look
> >> at /var/log/Xorg.0.log.old
> >> rather than Xorg.0.log). What's the gfx card?
> >> Thanks, I'll check that out, though I doubt that there'll be
> >> anything... I have intel 945. Another interesting thing is, I logged
> >> into a different account (which I use to testing/playing around) that
> >> currently has gnome-shell (as opposed to composited metacity) and so
> >> far after more than an hour no freeze. I'll see if it's still
running
> >> tomorrow.
> >>
> > Ok, that wasn't it. It froze in gnome-shell as well. In the exactly same
> > moment when external USB HDD was waking up (initiated by ls -l on some
> > folder in it). So over night I run full fsck on it. No errors reported,
> > but right now it seems to work (at least I wasn't able to reproduce the
> > freeze in single user mode, I'll see if it appears again in init 5).
> >
> > /var/log/Xorg.0.log* and /var/log/messages show no traces about the
> > freeze.
> >
> So it appeared again, after several hours of usage (so it's much
> better than it initially was), this time in gnome-screensaver. The USB
> drive appeared to be running, so I guess kernel was frozen as well.
>
So, after slowly eliminating possible causes, I came to (uncertain)
conclusion that f14 kernel does not like my network card (i.e. unless I
missed something, the only time I wasn't able to reproduce this yet is
without network cable plugged in). It also seems the freezes have got a
lot less frequent (but that might be purely coincidental).
Kernel:
kernel-2.6.35.4-28.fc14.i686
Network card:
06:01.0 Ethernet controller: Broadcom Corporation BCM4401-B0 100Base-TX
(rev 02)
Just FYI that I finally updated to a kernel that does not seem to suffer
from this (which is sadly in koji only):
$ uname -r
2.6.35.6-36.fc14.i686.PAE
$ uptime
22:11:38 up 3 days, 4:19, 10 users, load average: 1.40, 1.67, 1.80
Not sure if PAE could play any role in this -- I installed the PAE
version just for the heck of it (I don't have that much of memory to be
actually in need of the PAE)...
Martin