Re: [cobalt-developers] XTR Kernel

["William L. Thomson Jr." <support@xxxxxxxxxxxxxxxxxxxx>]

Ryan,

On Sat, 2002-12-21 at 12:18, Ryan Verner wrote:
> It seems the XTR is the most problematic
> Cobalt yet - and it certainly doesn't seem to be the hardware's fault, as
> they're very capable machines...

Yes, the XTR has had their fair share of problems. The hardware is good,
the only complaint I have is about the chipset of the ethernet
interfaces. It's a DP83815 chipset or something along those lines. The
RaQ4's, 550's, and LX50 all have Intel chipsets for the ethernet
interfaces. Not exactly sure about the 550. I prefer Intel chipsets to
all others. Also for some reason I have never been able to force the
speed and duplex settings of the XTR. I have tried, passing the
necessary params to the modules, but no go.

Now one thing I have always been curious about is the XTR's motherboard
having support for two proc's. I doubt Sun/Cobalt has any plans to
release a modified SMP kernel so I could toss in another proc.

So if I get my hands on another machine to use in place of my XTR. I
would like to experiment with the XTR. Like upgrading the ROM, (or what
ever it's called) to support 2.4 kernels. Then compile and install my
own 2.4 kernel with SMP support.

On a side note/question. What all does cobalt add/modify to the kernel?

I know there is
bandwidth module, don't really care about.
temp sensor, I very much care about
watchdog, mood issue somewhat important
LCD, to use the lcd along with the corresponding tools/binaries.

Am I missing something? I would assume all but the temp sensor and maybe
the LCD stuff to be universal to RaQ's?

The temp sensor, and possible LCD, would be specific to the XTR's
hardware?

Now back to the topic of Tomcat, I have never been able to replicate the
behavior of Tomcat on any other machine. I have a development machine,
with Tomcat on it that I use to develop and test apps before moving them
to the XTR. Now I have always assumed the problem to exist on both, and
the only reason it's not apparent on the development server is I can't
seem to simulate the actual real time load of the production server/env.

However that is an assumption and could totally be inaccurate. My gut is
starting to tell me that it's something specific to the XTR, or more so 
it's kernel?

It's hard to compare, because so much stuff on the XTR is out of date.
Not to mention the kernel.

I am sure there have been other bug fixes and etc to the 2.2 kernels,
thus the current version being 2.2.23. I compiled and installed my own
2.2.21 kernel for my Linux load balancing router, that has been up for
199 days straight. In that time there really should have been a kernel
update.

I find it humorous that an update is coming out, with another 2.2.16
kernel.

It's seem Cobalt is having problems with, or is not able to follow Sun's
philosophy of modularity. Like Sun ONE. Sun says take any part you do
not like out and drop in what you do like. Everything is modular.

It's seems Sun/Cobalt stuff is anything but modular and certain things
are specific. That's the only reasoning I can come up with. That or
Sun/Cobalt has a problem with keeping software up to date.

Like Apache, most of the world has moved on to 2.x, and while I
understand that is a big switch, why are we still stuck with a 1.3.20
distro of Apache?

Why are we also stuck with a 2.2.16 kernel?

Is the XTR now considered and EOL product? Why is it being treated as
such? Are the not selling? Are they just a big problem Sun/Cobalt would
like to go away?

Direction please, someone from Sun/Cobalt if your reading this, speak up
please.

Sincerely,
William L. Thomson Jr.
Support Group
Obsidian-Studios Inc.
439 Amber Way
Petaluma, Ca. 94952
Phone  707.766.9509
Fax    707.766.8989
http://www.obsidian-studios.com