I found this article (http://www.byte.com/documents/s=1794/byt20011107s0001/1112_moshe.html)interesting ... Any opinions?

cheers,
:beer:

they are both good OS's its all a matter of opinion really for which one you like.

I know Linux, especially Red Hat, gets a lot of negative comments on this board, but I find the 2.4 kernel to be very stable, and it appears to be able to match OSes that have traditionally been thought of as more robust.

but freebsd 4.5 is better :)

Benchmarks comparing Linux and FreeBSD has almost always been resulted in a relatively small margin. Small enough that it is not justified to switch to either o/s based only on these benchmarks. This benchmark is no exception.

With Windows however, the story is different :). If Linux or FreeBSD wins, people will start switching to Linux/FreeBSD. If Windows wins, people will start tweaking Linux/FreeBSD until it wins. :D

Originally posted by dektong
... Any opinions?


Opinions? Here??? :D


I am pretty near the end of a migration from Linux (Cobalt RaQ3 and RH 7.2 servers) to FreeBSD (4.4 and 4.5) for a couple for servers I run.

Performance-wise, I find FreeBSD and RH Linux 7.2 to be fairly close, with either one a little bit ahead depending upon what you are doing. The Linux kernel seems to handle multiple CPUs better. FreeBSD seems to handle file i/o a bit better. And so on...

Both OS's seems to be as robust as I need them to be, and both are significantly more reliable than Windows 2000.



So, if they are so close, why am I moving from RH Linux to FreeBSD 4.5? Well, FreeBSD "feels" like it is has a better underlying architecture and design. File placements, periodic cron jobs, hooks for script customization and software configuration are all, IMHO, better designed in FreeBSD.

As always, but especially in this case, YMMV.

Originally posted by dektong
I found this article (http://www.byte.com/documents/s=1794/byt20011107s0001/1112_moshe.html)interesting ... Any opinions?


Well, the maxusers setting is, umm, interesting. Hopefully now that 4.5-RELEASE has maxusers 0 = autoscale people like this will be able to run benchmarks somewhat more properly.

But generally I find it interesting that he compared the latest *stable* FreeBSD to the latest *unstable* Linux kernel. This is a symptom of the lack of proper development process: If you complain about a lack of performance with 2.2.x, you'll be told to use 2.4.x, but if you complain about (for example) having a new VM manager suddenly dropped in, you'll be told to go back to 2.2.x. This has, of course, improved somewhat now that 2.5.x has forked, but there still isn't the clear "this should absolutely *never* break" in 2.4.x the way there is in RELENG_4_5.