Mac Xserve to run Solaris X ? has boot bugs still
null
Just a question. Apple is doing testing with Solaris X for the Xserve but the
word is that there are still serious boot problems. The OSX gui is being
ported for the clients and who knows what else. Can we expect to see the OSX
gui on Solaris anytime soon?
null mac man
null@ksr_nospam_sports.com
Rich Teer
> Just a question. Apple is doing testing with Solaris X for the Xserve but the
> word is that there are still serious boot problems. The OSX gui is being
> ported for the clients and who knows what else. Can we expect to see the OSX
> gui on Solaris anytime soon?
I guess as the word is "out", I won't be violating my NDA by filling
in some of the details here.
Yes, that which was to be Solaris 10 is apparently being renamed
as Solaris X*. From what I can gather, Sun engineers have been
frustrated at the huge number of dependancies needed to build
GNOME 2.0, so they're planning to adopt Apple's Aqua GUI. I also
hear that Bill Joy wanted to get back to his BSD roots, and this
was a good way for him to do so. Obviously, the X moniker is a
nod to Mac OS X.
* Gotta love those imaginative marketroids!
As part of the cross-licensing agreement, Apple and Sun will be
basing their next generation of servers and workstations on the Hyena
(UltraSPARC V), and that's where the problems you allude to are.
I don't want to get my Sunsolve account revoked, so here's only
a snippet of the bug you're referring to:
Bug Id: 10403002
Category: kernel
Subcategory: other
State: verified
Synopsis: MP systems hang during multi-core boot
Description:
MP systems running on the Hyena quad core procs sometimes hang
when booting s10_b42. This is especially apparent in the
Chihuahua desktop.
The problem was first noted by Apple engineers, who were running
a port of Aqua on their Hyena simulations. The system hangs on
boot, after copying the miniroot.
[...]
s10_b42 was the first build of Solaris 10^H^HX that incorporated
Aqua. I think when the Hyena ships, Apple's servers will be based
on it, running Solaris X. Sun's next generation of workstations
(i.e., those after the US-IIIi ones that are due RSN) will be based
on a cut down version of the Hyena, presumably called US-Vi.
I presume that Sun will make Aqua available separately from Solaris X,
in much the same way that GNOME 2.0 isn't yet bundled with Solaris 9,
but is available for download.
--
Rich Teer, SCNA
President,
Rite Online Inc.
Voice: +1 (250) 979-1638
URL: http://www.rite-online.net
Sultanous
have a PowerPC port that was meant to run on the first macs or IBM
rs/6000's. The only concern are present architectural differences, and
as to how the base system libraries have to be rewritten in order to get
this to work on present day hardware.
This will allow apple to finnaly compete in the true unix realm and not
just claim to.
Sam
Dennis Clarke
>On Tue, 1 Apr 2003, null wrote:
>
>> Just a question. Apple is doing testing with Solaris X for the Xserve but the
>> word is that there are still serious boot problems. The OSX gui is being
>> ported for the clients and who knows what else. Can we expect to see the OSX
>> gui on Solaris anytime soon?
>
>I guess as the word is "out", I won't be violating my NDA by filling
>in some of the details here.
No big surprise about Aqua being leaked, sorry about the pun.
I think the idea was that Sun could address yet another unix like GUI and
then dtlogin would provide options for CDE, GNOME and Aqua with the Mac OSX
front end. The real deal is not the gui but the quad core processors that
are a joint project between Sun and Apple.
>
>Yes, that which was to be Solaris 10 is apparently being renamed
>as Solaris X*. From what I can gather, Sun engineers have been
>frustrated at the huge number of dependancies needed to build
Well .. GNOME is best on the UNIX Like world of Linux in the Intel world and
was a hell of a pile of problems to get to Sparc land where stability rules.
>GNOME 2.0, so they're planning to adopt Apple's Aqua GUI. I also
>hear that Bill Joy wanted to get back to his BSD roots, and this
>was a good way for him to do so. Obviously, the X moniker is a
>nod to Mac OS X.
>
>* Gotta love those imaginative marketroids!
Well Solaris X is pure 64-bit SVR4 still and a lot of old history Sparc will
fall by the way-side. It makes perfect sense that we can say goodbye to the
SBus days and hello to fast thin clients with OpenGL and slick Aqua front
ends. The idea is that the Apple Xserve units can provide super cheap speed
that comes at a per unit lower cost than the SunFire units. I think that Sun
came right out and siad that they wanted to address to lower end fast-cheap
market while they own the 64-bit super-computer end quite neatly with the big
SunFires. Where has the unix client gone? Aqua and Apple can fix that.
>
>As part of the cross-licensing agreement, Apple and Sun will be
>basing their next generation of servers and workstations on the Hyena
>(UltraSPARC V), and that's where the problems you allude to are.
> I don't want to get my Sunsolve account revoked, so here's only
>a snippet of the bug you're referring to:
>
>Bug Id: 10403002
> Category: kernel
> Subcategory: other
> State: verified
> Synopsis: MP systems hang during multi-core boot
> Description:
>MP systems running on the Hyena quad core procs sometimes hang
>when booting s10_b42. This is especially apparent in the
>Chihuahua desktop.
more on that ...
Initial quad core debugging on the MAC Xserve shows a thread timing issue
and a stable boot can only be achieved by disableing the other cores.
> $c
cpu_stop_self_c+0xDE(A0, D0, 0, E0, 0, 0)
> ::qpuinfo -v
ID ADDR FLG NRUN BSPL PRI RNRN KRNRN SWITCH THREAD PROC
0 0000140DE00 11 0 0 -1 no no t-137 2a10001fd40 (idle)
|
RUNNING <--+
ENABLE
ID ADDR FLG NRUN BSPL PRI RNRN KRNRN SWITCH THREAD PROC
1 0000142ff70 10 0 0 -1 no no t-53 2a100007d40 (sleep)
|
RUNNING <--+
DISABLE
ID ADDR FLG NRUN BSPL PRI RNRN KRNRN SWITCH THREAD PROC
2 300014e8000 1d 0 0 -1 no no t-0 2a10021bd40 (sleep)
|
RUNNING <--+
QUIESCED
DISABLE
ID ADDR FLG NRUN BSPL PRI RNRN KRNRN SWITCH THREAD PROC
3 30001588000 10 0 0 -1 no no t-12 2a10020fd40 (sleep)
|
RUNNING <--+
DISABLE
In any case the quad core units are being tested but only with a single core
running. The Apple and Sun people are almost certainly working on an Itanium
killer. If the SunFire is any indication it will need liquid cooling like the
old IBM 3090 mainframes except it will fit on your desk.
>
>The problem was first noted by Apple engineers, who were running
>a port of Aqua on their Hyena simulations. The system hangs on
>boot, after copying the miniroot.
>
No I think that they are up to a prototype boot with a single core.
based on the bug.
>[...]
>
>s10_b42 was the first build of Solaris 10
you mean X ... except it will be renamed or labeled badge or whatever
by the market droids. Who knows .. maybe we will end up with workstations
that run like SunFires in heat ... look like the Mac G4 and have an interface
called aqua that looks better than anything by Redmond MicroSquish.
>I presume that Sun will make Aqua available separately from Solaris X,
>in much the same way that GNOME 2.0 isn't yet bundled with Solaris 9,
>but is available for download.
who knows .. it may show up at Ximian. :)
I think that Bill should be worried about the number of killer OS's out there
that are sneaking up on the Windows client ... and probably eating it alive.
Dennis Clarke
Alan Coopersmith
|MP systems running on the Hyena quad core procs sometimes hang
|when booting s10_b42. This is especially apparent in the
|Chihuahua desktop.
You know, I never have liked that code name - after all the Mexican food
code names, it always seemed like a wholly wrong spin on "eating your
own dog food."
--
________________________________________________________________________
Alan Coopersmith al...@alum.calberkeley.org
http://www.CSUA.Berkeley.EDU/~alanc/ aka: Alan.Coo...@Sun.COM
Working for, but definitely not speaking for, Sun Microsystems, Inc.