[N3WBI3]

There was a kernel bug in Redhat which if you had a database server or other hog with insufficient memory kswapd would go nuts. You would lose the box for 15-20 seconds, nothing was logged. It was easy enough to fix by using sysctrl.conf but you took a hit in performance for it.

Solaris is a great solid OS, and I would buy Suns hardware running it any day but they have not shown enough long term commitment to the x86 platform for me.

Solaris was the better OS pre kernel 2.4, but its so close to a draw now I concern myself more with the upgrade paths than the performance..

[Boundless]

> ... I would buy Suns hardware running it any day but
> they have not shown enough long term commitment to
> the x86 platform for me.

Well, they seem to have pretty much bet the company on it.
Or did you mean that their track record on x86 is short?
The latter is certainly the case.

If Sun isn't committed to x86 long term, then there will
be no Sun long term. Sun's strategy may fail, but I don't
see any abandonment of x86.

Heck, even Apple is on x86 now, and it's unlikely that
they will jump to another ISA anytime soon. I'm no
advocate for x86, but the message is pretty clear here
(except, apparently, to the remaining passengers on the
good ship Itanic :-)

[Prost1]

An ancedote...
I was standing next to Bill Joy after a presentation (circa 1990) and someone asked a question. He answered that he was thinking about rewriting Sun's OS Kernel, as soon as he found the time...This was BSD4, I think. Definitely pre-Solaris.

Now, I'm not an engineer of any type, but to think, "I'm thinking about rewriting the Kernel..."

Is that complex, simplex or amazing?