On Wed, Jan 05, 2000 at 03:19:44PM +0100, Thomas Michael Wanka wrote:
> Hi,
> 
> I cannot deny bugs in Linux and I donīt want to. As You know the 
> power down problem could also be caused by the mail server or 
> other software and this problem may allready be solved.

Time for one of "David's famous simplistic examples" for people to
criticize ;-)

If, as a programmer, I write a (pseudo)script that says:

touch lockfile
mv file1 file2
rm lockfile

I want to be pretty sure that 'mv' finished before I'm clearing
the lock. Under FFS (NetBSD) this is guaranteed true, under EXT2FS
(Linux) it is not.

At this point, no matter how bright the application programmer is with
his mail server, he doesn't get underlying safety - (unless he writes
a bunch of Linux specific work-around code) even though he's doing
things the 'right' way.

If you're interested in the details, this book provides a good overview
of the issues.

Practical File System Design with the Be File System 
http://www.amazon.com/exec/obidos/ASIN/1558604979/beincorporated/104-1705366-3762852

> But even if these problems are to be expected to still exist, if You 
> need to run programms only available for linux or have hardware not 
> supported by NetBSD, would You really recommend NetBSD? 

Sometimes, yes. At a previous job, a ran a couple machines worth of
Quake servers, under NetBSD. (Only Linux binaries available...)
Each machine ran 3-4 servers, and people were amazed at the performance,
and availability - compared to other Quake servers.

(This was at a cable company, lots of people using rate 10000, so they
weren't treating the server lightly.)

-- 
David Maxwell, david@vex.net|david@maxwell.net -->
All this stuff in twice the space would only look half as bad!
					      - me