Not sure if this is related but here is a bugzilla entry that gives some
items to change to try to fix this issue.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=170008
-Connie Sieh
On Tue, 11 Oct 2005, John Franks
wrote:
> I can confirm this on some optiplex gx280's that I have.
>
>
> Here is what I observe:
>
> 1. Same behavior on genuine RHEL4 (no surprise)
>
> 2. No problem with non-smp kernel.
>
> 3. Apparently no problem when booting into runlevel 3 i.e. no X.
>
> 4. With smp kernel, removing "quiet rhgb" from the end of the grub boot
> line allows booting and X to start. However, after first logout it
> hangs when gdb tries to restart X.
>
> For me also the workaround seems to be the non-smp kernel. I haven't
> tried to search bugzilla yet; there may be something there.
>
>
>
> On Tue, 2005-10-11 at 09:12 -0500, Gregory Harris wrote:
> > Folks,
> >
> > A couple of interesting things to note:
> >
> > (1) I can boot the 2.6.9-22 kernel just fine. It is just the
> > 2.6.9-22.smp kernel that has problems.
> >
> > (2) When booting the 2.6.9-22.smp kernel, the system starts up, and I
> > can ping the system UNTIL it leaves the graphical boot screen to go back
> > to the console to start gdm for logins. Then the gdm screen is blank,
> > there is a flashing cursor up in the uppper-left-hand corner, and the
> > system immediately stops responding to pings and the keyboard/mouse are
> > locked. The caps/num/scrolllock lights turn blank, and pressing the
> > numlock/capslock/scrolllock have no effect on their respective lights on
> > the keyboard.
> >
> > (3) If I enter the bios of these systems, and turn off the Exec Disable
> > memory protection, the system does boot up normally on the 2.6.9-22.smp
> > kernel, but does take a good deal of time to start the IIMF input server.
> >
> > So at least for now it appears the workaround is to either run the non
> > smp kernel, or turn off the exec disable memory protection in the bios.
> >
> > I know this new kernel did fix security problems with memory access on
> > multiple architectures. It is possible the developers did not account
> > for the exec disable option in the bios. Is this a 'bad' thing to have
> > turned on anyhow?
> >
> > Gregory Harris wrote:
> > > Nope, just plain OptiPlex GX280's. Integrated graphics cards,
> > > integrated NICs, integrated sound. All of them have the DVI conversion
> > > adapter installed on them though, which the flat panels use but the
> > > CRTs don't. It seems to take forever to boot, but the network adapter
> > > does get brought up but I can' t ping the systems remotely.
> > >
> > > On Oct 10, 2005, at 5:19 PM, Connie Sieh wrote:
> > >
> > >> Gregory,
> > >>
> > >> Any "special" hardware on this Dell?
> > >>
> > >> -Connie Sieh
> > >>
> > >> On Mon, 10 Oct 2005, Gregory Harris
> > >> wrote:
> > >>
> > >>
> > >>> Yeah, I felt dumb after I wrote that because I forgot about grub.
> > >>> Anyhow, it's a rather interesting problem that a kernel update would
> > >>> cause this much havoc.
> > >>>
> > >>> Thanks,
> > >>> Greg
> > >>>
> > >>> On Oct 10, 2005, at 4:51 PM, Connie Sieh wrote:
> > >>>
> > >>>
> > >>>> Gregory,
> > >>>>
> > >>>> Grub should give you the option to boot the older kernel.
> > >>>>
> > >>>> -Connie Sieh
> > >>>>
> > >>>> On Mon, 10 Oct
> > >>>> 2005, Gregory Harris wrote:
> > >>>>
> > >>>>
> > >>>>
> > >>>>> Anybody else having serious problems with the 2.6.9-22 kernel that
> > >>>>> just came out for SL4?
> > >>>>>
> > >>>>> I have yum set to autoupdate. I have just found out that all of the
> > >>>>> 30 or so Dell Optiplex GX280's we have that have been rebooted are
> > >>>>> all hanging on boot. There doesn't seem to be anyway to get to a
> > >>>>> root console on them either, it doesn't bring up networking and hangs
> > >>>>> on the "setting up swap space" or it goes to a blank screen.
> > >>>>>
> > >>>>> At this point I don't see anyway to revert to the previous state
> > >>>>> other than reinstalling. It's not a big deal here though since we
> > >>>>> use network home directories. Just want to get the word out before
> > >>>>> it bites someone else.
> >
>
|