Re: question about logging kernel problems when kernel blows up

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

 



There are other (mostly simpler) ways to get some commonly
needed error output from the kernel, unmodified, as well.  
Standard system logs are often useful, and should be
consulted first, such as /var/log/messages .  If your system
is not completely dead, the "dmesg" utility could help (but
may only deliver what is in the logs anyway). And you can
possibly get the kernel to dump some info through the "magic
SysRq keys", even after a kernel panic, if this feature is
activated in advance; see
/usr/share/doc/kernel-doc-2.4.20/sysrq.txt, or where-ever
"locate" shows that file lives in your distribution, for
details.  To activate it, put the following line in
/etc/sysctl.conf:

kernel.sysrq = 1

Don't forget to manually activate it the first
time (automatic on bootup).

And you might be able to get the stock kernels to
deliver even more diagnostic info by turning on
features related to that using the sysctl utility
(consult the man page -- I haven't checked).

Mostly, recompiling kernels for various features
is out-of-date strategy.  Stock kernels now
normally have everything a distributor can
reasonably bundle with it, and you only have to
figure out how to activate those features, such as
tuning the kernel or loading modules (often
automatic).  For exceptions to that, there are web
archives where you can get other stuff: for
instance, low latency kernels with the advanced
alsa sound drivers (for my case), already built in
are available, for professional audio work or just
better performance and features in your card.
Other kernels are out there for pretty much
whatever else you need.

Also, much info is stored in the /proc (fake)
filesystem directories, and can be conveniently
accessed through front end utilities like lsdev.

And a search at http://www.google.com/linux can be
very helpful.  And there are yet more resources,
such as HOWTOs (see the blinux FAQ).

One problem with helping you with a concise answer
is that you have been very vague about the actual
nature of the problem; otherwise someone here
might have been able to tell you exactly where to
get the info you need.  And a common strategy with
vague postings is to ignore them, thus protecting
one from info overload (and related hassles and
drains) on the net (I do it all the time).

LCR

On Tue, 27 Jan 2004, Andor Demarteau wrote:

> On Mon, 26 Jan 2004, [iso-8859-15] Sébastien FRANÇOIS wrote:

>  > > I have a friend who is trying to help me
>  > > debug a kernel problem.  I am wondering if
>  > > there is any way to log the problems that
>  > > the kernel encounters when you load it and
>  > > it blows up.  He needs this to help me
>  > > debug what is wrong.  Thanks for the help.
>  > There is a way to request the kernel to send
>  > info over a serial line, I have never done
>  > this though
> it's an option in the kernel itself caleld
> "output to serial dump terminal" it effectively
> gives all kernel-output over the serial-poort. A
> bootoption controls which port is used.  You can
> also login via the serial terminal, but you have
> to uncomment the correct line in /etc/inittab as
> well to make this possible.


-- 
L. C. Robinson
reply to no_spam+munged_lcr@xxxxxxxxxxxxxxxxxxx

People buy MicroShaft for compatibility, but get
incompatibility and instability instead.  This is
award winning "innovation".  Find out how MS holds
your data hostage with "The *Lens*"; see
"CyberSnare" at
http://www.netaction.org/msoft/cybersnare.html


_______________________________________________

Blinux-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/blinux-list

[Index of Archives]     [Linux Speakup]     [Fedora]     [Linux Kernel]     [Yosemite News]     [Big List of Linux Books]