[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: kernel messages (from a newbie)



On Tue, Nov 26, 2002 at 02:49:56AM +0000, Pedro Jorge wrote:
> Hi,
> 
> I'm a newbie to Debian. I've been using 'commercial' linux (Red Hat and
> Suse) in my desktop for some time, but i've never had a real problem
> with them, so i don't know much about the OS.

Commercial Debian variants include Xandros (with WINE added) and
Libranet (a Canadian company;  nice K desktop setup they have there).

> I bought a Presario 920, with an Athlon XP 1.67MHz, and I couldn't
> install Red Hat 8.0. So i've decided it was time to try Debian, and
> make myself learn some more about linux.
> 
> I've got the system - woody 3.0 -  running (without sound, nor modem, 
> for now),
> but it prints a kernel's message lots of times:
> 1.    'kernel: neighbour table overflow'

This means localhost is not setup properly.  Go to
/etc/network/interfaces and make sure that you have
	auto lo

and a stanza for the lo device giving it the 127.0.0.0 network, class C.
(or A, if you're really strange but expect to use it.  Just *do* use the
matching netmask.)

> It keeps running, aparently normaly.
> Can anyone help me, and tell me what this means?

It's the arp cache overflowing because it cannot tell that 127.0.0.1
belongs to "itself" and keeps asking via ARP "hey who has this address"
-- of course it gets immeidate answers, but the little cache floods.

Or at least that's as far as I can tell.  Anyways it's always localhost
that's missing.  I've not seen that for other interfaces, even in flood.

> There are someother messages from the kernel (in /var/log/messages)
> that i don't understand:
> 2.    'mtrr: no more MTRRs available'
 
memory type range register (if I recall correctly) are used by X and a
few other apps.  You either have them or you don't;  a system never has
huge gobs of them.  So they are bound to run out occasionally.

It's harmless.  Having no MTRRs at all still works too, it's just a 
smidge slower for some things.  You have to have a rather modern CPU 
to have any to use.

> 3.    ' kernel: probable hardware bug: clock timer configuration lost
> - probably a VIA686a.
>     kernel: probable hardware bug: restoring chip configuration.'

Possibly a good result of "good/bad firmware, works in progress" marked 
during the kernel build.  It thinks it found something iffy, which it
can fix.
 
> Do this ones have anything to do with the first?

no.

> I belive this last one is because the temperature sensor, right? I've seen
> some notes on lm-sensors, but i'm a little afraid to start messing around
> without some help.
 
lmsensors is the userland app that deals with temperature checking;
that looked like a kernel complaint, not a userland one.  And it said
clock timer.  Nope, I don't think that's a thermal alarm.  But did you
hear the system change at all to go with that?  e.g. the fan turn on or
go double speed, the screen get dimmer, etc.
 
> With my regards,
> Pedro
> 
> PS- I promise i'll keep next messages short :^)

Don't try *too* hard, it's not always right to clip out good info.  
Although saving config attachments for later or just sections to ask
about is usually best.

* Heather Stern * star@ many places...
                * Starshine Technical Services -*- 800 938 4078



Reply to: