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.

No the mail server is ok (it's sendmail) and it is clearly a problem with
linux's filesystem.

> 
> To the bad block problem, I have senn such problem only on my NT 
> system with a defective Jaz medium, I guess such a problem was 
> very soon be detected and reported, so You can expect a patch 
> allready exists to solve this problem.

This existed in the 2.0 kernels and is still here in 2.2. I think this comes
in a general attitude of the linux developers: "this is not a normal
situation".

> 
> Such things are bugs, and should not be part of the evaluation of an 
> OS. If You could say, that the problems reported by You are not 
> caused by bugs in programms but the OS itself and still exist after a 
> couple of months after the report to the programmers, THEN it was 
> good reason to include it in the evaluation of an OS!

I'm not sure it is considered as bugs by the linux community, at last the
filesystem problem is not, for sure.

> 
> 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? 

Hardware: I buy hardware for applications. For a critical system I wouldn't
chose the OS because a ethernet or SCSI board is not supported by NetBSD,
I would just buy another one.
Software: if it's really needed and only available for linux of course I would
install a linux system for it, but this should be really marginal (I don't know
of a program available only for linux). For the average commercial soft
I'd use solaris instead.
> 
> I use several machines with different OS and I currently would not 
> want to replace any of the OS (good MPS support from NetBSD and 
> a firmly tested VMWare for it would change that probably soon ;-)).

Mee to ! My dataless workstations are running linux (because of staroffice,
and a few other things). If one goes bad I just reinstall it: I have
kickstart set up, I just need to put the floppy in, hit reset, wait for the
kernel to load and remove the floppy. 1/2 hour after it's back up and running.
Clearly this is not acceptable for a server.

PS: another thing I forgot to add: my linux clients have problems with the
solaris NFS server, sometime they just hang, sometime the local cache gets
corrupt. The ethernet driver is a bit flacky for these boards (SMC) but
this shoudl't prevent NFS from recovering.

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--