Re: [tlug] FreeBSD.....Linux what's the difference

[Scott <scottro@example.com>]

On 10:43 2002/05/02 +0900, Matt Doughty wrote
> >
> > Linux seems to do better on iffy hardware.  Probably (hopefully?) not an
> > issue on a server, especially a mission critical one, but for example, 
> on a
> > dying hard drive, Linux and MS would still install but FreeBSD would fail
> > to install. (I'm only experienced with FreeBSD, so when I say BSD, please
> > realize that's what I mean)
>
>This would be a 'good thing'(tm). I don't want to install an OS on a HD that
>is later going fail and take my data with it. So chalk up a point for *BSD.


That wasn't a negative. This drive actually came up on this list--I asked 
for opinions  as to why something wasn't working, kinput2, I think, the 
next time I booted the thing, I seemed to have lost ping, etc. etc.---the 
drive gradually gave up the ghost, but it died a slow death.

> >
> > KVM switches--which might be an issue--BSD doesn't deal with them well if
> > you have to run X for any reason.  The mouse won't work, if you have to do
> > any point and wimp stuff (he says, writing in Windows--hey, I spent the
> > last few days on distro installing binges.)  :)
>
>I have had more KVM issues with linux actually so YMMV.

No kidding.  Now that's interesting---I've never had trouble with Linux and 
KVM's.  With FreeBSD (I repeat, that's the only one I've played with for 
more than a day) there is a known bug--there were various suggestions, 
recompiling the kernel and setting flags, but it didn't work for me.  A 
later search on deja indicated that the gentleman working on it was about 
to throw up his hands in disgust--this was several months ago, however.


> > Much of this may not be an issue on a server, but might be if you're
> > playing with it at home before putting it on the server.  Also, if playing
> > with it at home first, be sure that, if using IDE, that you have shielded
> > 80 pin ATA100 (at least) IDE cable or you'll get a bunch of errors on boot
> > (though it will work without problem)
> >
>
>But the question is whether that is because of a problem in the BSD driver
>or are the linux/windows drivers either not seeing or just ignoring the
>errors? If the drive works fine anyway then there certainly is no idication
>that the BSD driver is functionally inferior. It might just be either more
>informative or more thourough in its testing.


Exactly---hope I wasn't giving the impression that I considered this to be 
a defect.  It can be an early warning system.  On at least one of the boxes 
where I had this issue, a closer look at dmesg indicated that Linux was 
giving me CRC errors, which also disappeared when I spent the 3 dollars for 
good IDE cable.  (I don't think I checked after that, because now I put in 
good cable first thing).


FreeBSD seems (this is subjective, I haven't done any benchmarking) to be 
faster, but actually, a couple of smaller Linux distros, Gentoo and Arch 
Linux to be precise, seem to be close. Gentoo's portage is similar to the 
FreeBSD ports--just that as Gentoo is rather new, it still seems a bit 
buggy, and although some people are doing it, I'd be leery about putting it 
on a mission critical machine at this point.

Anyway, aside from the KVM issue, which isn't that big a deal to the home 
user, I was mentioning these other points more for the home user---and, I 
wouldn't consider them defects--it's just that FreeBSD does seem more 
sensitive to iffy hardware--as you say, this is not a Bad Thing (TM).  (The 
ASUS MB issue was a defect, but the maintainer of the PR took the trouble, 
when I dropped him an email asking about progress, to send me the quick and 
dirty hack to get it working, which I thought was kind of him)

Re the cables--FreeBSD showed me the errors, then I began examining my 
Linux dmesg more quickly---hrrm, I said, CRC errors, that is not a Good 
Thing (TM).  One wonders what was going on behind the scenes as my Win2k 
booted up.  :) If i gave the impression that I thought this was a defect, I 
apologize.

Scott Robbins