On Fri, 29 May 2009 16:55:24 -0700, Steven Schlansker <s...@eecs.berkeley.edu> 
wrote:
> Hm.  I was actually under the impression that you wouldn't gain much  
> by compiling your own kernel (except for maybe some disk space). 

No. Compiling one's own kernel allows for some definitions,
e. g. having only the support in the kernel for the hardware
that is actually existing in the specific settings. Furthermore,
in some cases variables need to be set at kernel compile time.

For example, in FreeBSD 5 the only way to get IPFW was to
compile it into the kernel. Today, there's a module for this.

As you mentioned correctly, speed considerations may also lead
you to the requirement of compiling the kernel.

Speaking for myself, I like to have a custom kernel on my
own machine (which is somewhat special, so it deserves it),
but on customers' systems, going with "as most untouched as
possible" is often the way I choose. This allows me to safely
use tools like freebsd-update.



> Is there a strong reason to compile your own  
> kernel for "production" machines? 

In most cases, production machines do not require compiling a
kernel. But it depends on the setting - if you essentially need
something that cannot be done via kernel tunables or loadable
modules.



> The discussion online is not  
> conclusive (then again I'll probably just get contradictory opinions  
> again here!)

Of course. :-)





-- 
Polytropon
>From Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...
_______________________________________________
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Reply via email to