Re: [Testbed-admins] GENI proposal

[Michael Blodgett <mblodget@cs.wisc.edu>]

Hello Jay, we have a number of sun 4200 an earlier x86 Sun machine that 
we used outside of our Emulab testbed here at Wisconsin.  I'm also 
waiting for an order of 2250 to arrive to our department IT group, which
I was going to test with before hopefully purchasing some additional 
nodes for our testbed.

I put a 4200 on our testbed, it looks like there is an issue with 
FreeBSD and 4G of memory, I get a kernel panic on boot see below. If
I drop the memory down to 2GB it will boot.  I think it has something
to do with how the mfs_root is mapped into memory though that's just
a guess from some newsposts I found through google.

Mike

Enter boot spec: loader:/tftpboot/freebsd 

Loading /boot/defaults/loader.conf 

/boot/kernel text=0x5d59bc data=0xa4bc0+0x5fd4c 
syms=[0x4+0x72d00+0x4+0x8ef01]
/boot/acpi.ko text=0x44e9c data=0x24e0+0x1b8c 
syms=[0x4+0x7dd0+0x4+0xab6e]


Hit [Enter] to boot immediately, or any other key for command prompt. 

Booting [/boot/kernel]... 

KDB: debugger backends: ddb 
       id=0 auid=4294967295 subj=system_u:system_r:hwclock_t:KDB: 
current backend: ddbrademark of The FreeBSD Foundation.?, addr=?, 
terminal=console res=success)'
Copyright (c) 1992-2008 The FreeBSD Project.30 MDT 2008
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 
1994TESTBED-PXEBOOT-GENERIC
Turning The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 6.4-PRERELEASE #0: Thu Sep  4 08:49:30 MDT 2008

root@node.freebsd63-node.emulab-ops.emulab.net:/usr/obj/usr/src/sys/TESTBED-PXEBOOT-GENERIC
fault codeng all md devi= supervisor write, page not present
instruction pointer for = 0x20:0xc0928dc8.1 disabled
Fatal trap 12: page fault while in kernel modesabled
fault virtual address   = 0xbff1e00020d0c.1 disabled
fault codentterrupt for = supervisor write, page not present
instruction pointer     = 0x20:0xc0928dc8def32 1, gran 1
stack pointerags        = 0x28:0xc3020d00ed, resume, IOPL = 0
frame pointerss         = 0x28:0xc3020d0c
code segment0 tid 0 ]   = base 0x0, limit 0xfffff, type 0x1b
Stopped at      pmap_map= DPL 0, pres 1, def32 1, gran 1,4)
processor eflags        = interrupt enabled, resume, IOPL = 0
current process         = 0 ()
[thread pid 0 tid 0 ]
Stopped at      pmap_map+0x28:  movl    %eax,PTmap(,%edx,4)



Robert P Ricci wrote:
> The best place to send these sorts of questions is the testbed-admins
> list, which I've CCed.
>
> We don't have any experience with Sun machines at all. I'm not sure that
> anyone is running an Emulab based on them. What we (and, I think, most
> others) use are typically Dell server-class machines.
>
> I assume these are x86-based Sun machines? They *might* work, but you
> will need to verify that they can be booted with PXE.
>
> Our page discussing hardware recommendations is here:
>     http://users.emulab.net/trac/emulab/wiki/HWRecommend
>
> Thus spake Seung-Jong Park on Tue, Apr 28, 2009 at 02:24:23PM -0500:
> > Robert,
> >
> > I have a question regarding servers which we are going to purchase.
> >
> > We will purchase SUN server 4240.
> > We concern that SUN 4240 has some compatibility issues with FreeBSD 6.3 or 
> > earlier version which Emulab requres for testbed.
> >
> > Do you think that this is a real problem ?
> >
> > Thanks for your help
> >
> > Jay
> > .
> > Seung-Jong Park, Ph.D.
> > Assistant Professor
> > Computer Science and CCT
> > Louisiana State University