[ale] Weird Kernel Behaviour

James P. Kinney III jkinney at localnetsolutions.com
Mon Nov 12 15:40:26 EST 2007


On Mon, 2007-11-12 at 14:53 -0500, Calvin Harrigan wrote:
> James P. Kinney III wrote:
> > On Mon, 2007-11-12 at 13:31 -0500, Sergio Chaves wrote:
> >> I think I have a good one here:
> >>
> >> Dell Latitude D810
> >> RHEL 4, fully updated
> >>
> >>  I've been using  the current Kernel, 2.6.9-55.0.12.EL, for a week or so now.
> >>  Out of the blue, today,  I lost networking and could not reconnect.
> >>  Rebooted system and now the Kernel does not see the LVM.
> >>  Booted on old Kernel and everything is fine. Clues anyone? 
> > 
> > Hard drive horked the 2.6.9-55 kernel. Likely caused by the power surge
> > that broke your networking session.
> >>    
> 
> Ummm...  The D810 is a laptop...

Makes for a more fragile hard drive. Power surges can happen over any
access port - power, network, modem, USB. If a device can be physically
connected, that can be an entry point for a surge. The surge can be
static induced (most systems can now dissipate some static buildup on a
usb device - that's why the outer metal shield is ground.)

I have seen power surges come in through a running network line. Wiring
fault causes a cat5 line to be too close to a high current device. That
device starts up (think motor) and the resulting induced voltage in the
adjacent cat5 line cooks a NIC. Faraday's Law is verified again.

Extreme end of this example is a pair of near-miss lightning strikes I
have experienced. One hit a tree two door down and the surge was induced
in the phone line 300 feet away. That line was connected to my modem. My
machine was trashed completely (hard drive heads were physically welded
to the platter!! Example 2: tree in neighbors yard was hit and the
strike jumped to a tree in my yard. That tree is less than 10 feet from
the exterior wall of my office. I was just inside that wall when this
occurred. The extreme electric field lit up a pair of unopened
fluorescent light tubes so bright the shrink wrap fused to the glass.
Every device connected to a cat5 line in the office (and house!) was
destroyed instantly - switches, routers, hubs, nics. Over the next 2
weeks all hard drives and controller cards failed in every machine that
had a damaged nic. Over the next 2 months I was plagued with gremlins
until all systems were completely replaced to eliminate the motherboards
and CPU and RAM. The monitors and UPS devices held up fine. Note that
the strike was not to a power line or building. It was just nearby.

It's pretty easy to generate an electric field high enough to induce a
50V spike down a network line. 50V makes a dead nic. 30V will make a
network connection hiccup and may or may not cause damage.
-- 
James P. Kinney III          
CEO & Director of Engineering 
Local Net Solutions,LLC        
770-493-8244                    
http://www.localnetsolutions.com

GPG ID: 829C6CA7 James P. Kinney III (M.S. Physics)
<jkinney at localnetsolutions.com>
Fingerprint = 3C9E 6366 54FC A3FE BA4D 0659 6190 ADC3 829C 6CA7
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part




More information about the Ale mailing list