Booting Solaris on OpenSPARC

Hi all,
I'm trying to boot Solaris on a Xilinx ml411 board with one OpenSPARC core.
I wrote the following files to the matching OpenSPARC address.
0xff f000 0000 - OpenSPARCT1_Arch.1.5/hypervisor/src/greatlakes/ontario/t1_fpga/prom.bin
0xff f008 0000 - OpenSPARCT1_Arch.1.5/S10image/openboot.bin
Can someone explain the error below.
Thanks,
Teck Hua
<Output>
Entering hypervisor
Scrubbing initial RAM
Copying from ROM to RAM
Entering hypervisor
Running from RAM
Hypervisor version: Hypervisor 0.1 2007/08/03 11:23 [greatlakes:ontario:t1_fpga] Not scrubbing remaining hypervisor RAM. Using pre-initialized memory.
Not scrubbing guest memory. Using pre-initialized guest memory.
Using pre-initialized Guest stand-alone program memory image.
WATCHDOG: pcpu: 00000000 tl: 00000003 tt: 00000010 gl: 00000003 trap state:
tl: 00000001 tt: 000001ff htstate: 00000004 tstate: 0000000000000400
tpc: fffffffffffffffc tnpc: fffffffffffffffc
tl: 00000002 tt: 000001ff htstate: 00000000 tstate: 0000000814001400
tpc: fffffffffffffffc tnpc: fffffffffffffffc
tl: 00000003 tt: 00000010 htstate: 00000000 tstate: 0000024400001400
tpc: 0000000000800020 tnpc: 0000000000800024 trap globals:
gl: 00000000
%g0-%g3: 0000000000000000 0000000000000014 0000000000001890 0000000000070000
%g4-%g7: 0000000000000000 0000000000064dc8 0000000000064dc8 0000000000009ab8
gl: 00000001
%g0-%g3: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
%g4-%g7: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
gl: 00000002
%g0-%g3: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
%g4-%g7: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
rtba: 0000000000800000

Hi,
OpenSPARC T1 v1.5 does not support Solaris boot. This feature is scheduled for v1.6 which is expected to come out soon.
With v1.5, you can boot hypervisor and can run stand-alone C program. Method to link C program executable in the hypervisor binary is described in the Design and Verification document.
openboot.bin included in the S10image directory is meant for booting Solaris on Legion full-system simulator.
HTH.

Similar Messages

Problem booting Solaris 10 on ProLiant server after installation.

Hi all!
I have a brand new HP ProLiant DL360 G5 server. It contains 6 internal
hard drives. RAID 5 is configured on these drives using built-in HP
Smart Array P400i controller. I've just installed Solaris 10 8/07 on this
server. Solaris requires a driver for P400i controller. I'm using Driver
Update CD created from HP provided ISO image. Solaris 10 is
officially supported OS for DL360 G5 server.
Installation proceeds as expected:
1) At the begining of Solaris installation, Driver Update is used to
load HP driver for raid controller. During installation, disk is
recognized as c1d0t0 (the only disk in system); one partition is
created on this disk.
2) Disk is devided to 6 slices:
c1d0t0s0 - swap, 4GB
c1d0t0s1 - /, cca 32GB
c1d0t0s[3,4,5,6] - /zones/[a,b,c,d] - each cca 160GB
3) At the end of installation, CPQary3 driver is installed to target
system without any errors reported. No errors are reported during
installation at all.
After reboot, GRUB is started. It shows boot menu containing two
options:
- Solaris 'normal' boot
- failsafe boot
PROBLEM: Selecting 'normal' boot leads to immediate system restart.
According to documentation, this case occurs when PCQary3
driver is not installed.
I'm able to boot using 'failsafe' option in grub. After entering single
user, I load PCQary3 driver to runtime using 'install-du' command.
I'm also able to find instaled Solaris using 'install-recovery' and mount
its root volume to /a directory. Necessary driver IS installed inside
/a/kernel/drv/ directory. Options 'bootpath' in /a/boot/solaris/bootenv.rc
file is set up correctly pointing to device associated with /dev/dsk/c1d0t0s1
device. Also GRUB is able to recognize UFS filesystem for boot partition.
I've already tried (without success):
1) Use RAID 1+0 instead of RAID5
2) Create only 3,5GB partition instead of 700GB large one.
3) Place / on c1d0t0s0 instead of c1d0t0s1.
4) Re-create boot archive using 'bootadm update-archive -R /a'
So, everything seems to be installed correctly, but system cannot boot.
Do you have any ideas how to fix this problem?
Jozef

Sounds like an issue I just had with a brand new X4150, installation and failsafe worked fine, but normal booting led to immediate reboot of the server. This was due to an incompatibility of solaris and intel's newest generation xeon chips.
You can still boot 'normal' solaris in 32-bit mode:
hit 'e' at the grub menu, 'e' again to edit the 'normal' boot entry and add 'kernel/unix' after multiboot. This should boot up solaris in 32-bit mode. From there, download and install the latest patch set (125370-06 and 127112-05 are fixing this issue). After that you should be fine.
Bjoern

Dual boot Solaris 8 and Windows XP

Hello. I currently have windows xp as my OS. I would like to install Solaris 8 as a secondary OS. I have a separate hard drive I would like to put Solaris on. My question is, will the solaris installation allow me to choose which drive to install solaris to? Also, how can I set up my system to boot windows if I do nothing, but boot solaris if I choose to? Thanks for any help, I'm new to solaris and to dual-booting.

Try this URL:
http://sun.pmbc.com/faq/9.html
Hope it helps.
Senthilkumar
Developer Technical Support
Sun Microsystems, Inc.
http://www.sun.com/developers/support

How do can I use Jumpstart to boot Solaris from Slave Drive

Hi all,
I have XP on Master Drive and I used Jumpstart to install Solaris on Slave Drive.
This is the profile:
install_type     initial_install
system_type     standalone
usedisk          c0d1
boot_device      c0d1 update
root_device     c0d1s0
fdisk           c0d1 solaris all
partitioning     explicit
filesys          c0d1s0 10000 /
filesys          c0d1s1 1111 swap
filesys          c0d1s7 free /export/home
cluster          SUNWCreq
Now, I specified that the boot_device is c0d1 but after the installation the boot device is c0d0 which is my Master Drive...
...when I use Windows Boot Manager to boot Solaris... it cannot find the boot files on the master drive so it runs the DCA...
... so under the DCA I have to tell Solaris that the boot files are in my Slave Drive then I save it.
...I don't want to make this change everytime
any ideas that would make this install hands free?

...don't know the answer to that but here is an alternative:
... use this as your finish script
... it changes the bootpath in bootenv.rc to point to the slave drive
#!/usr/bin/perl
open(INFILE,"/a/boot/solaris/bootenv.rc");
open(OUTFILE,">/a/boot/solaris/bootenv.rc_out");
while(<INFILE>) {
chomp;
#--cmdk@0,0 is the master drive
#--cmdk@1,0 is the slave drive
s/cmdk\@0\,0/cmdk\@1\,0/g;
print OUTFILE "$_\n";
#...replace bootenv.rc_out with bootenv.rc
...happy programming

Can't boot Solaris 10 during installation

I am stuck up with a petty problem while booting Solaris 10 1/06 from the installation CD no.1. The GRUB boot loader appears fine, but when I try to boot the Solaris kernel and module, a message appears
Error 28: Selected item cannot fit into memory
My x86 computer's configuration is as follows:
Intel Pentium 4 processor 2.83 GHz
RAM 128 MB
HD 80GB
I can't figure out whether the problem is really due to insufficient memory or do I need to install more RAM on my platform.Can someone help me out with this?

128mb is plenty of memory. No it isn't plenty of memory. Solaris 10 Update 1 or Solaris Express needs at least
256mb of memory. The initial root ramdisk that is loaded from the CD/DVD
uncompresses to ~ 128mb already, leaving no space for the SunOS kernel on a
system with only 128mb of memory!

Cannot boot Solaris 11.1 Live Media for x86

Hi,
if I boot from Live-CD (burned "sol-11_1-live-x86.iso"), it comes:
---begin---
atapi_id_upate: select failed
ata_command: select failed DRY 0x1 CMD 0xef F 0x66 N 0x0 S 0x0 H 0x0 CL 0x0 CH 0x0
WARNING: /pci@0,0/pci1002,43a0@15/pci-ide@0/ide@1 (ata1):
timeout: abort request, target=1 lun=0
---end---
Any hints?
Thanks!
dynup
Edited by: dynup on 19.02.2013 02:42

dynup wrote:
---begin---
atapi_id_upate: select failed
ata_command: select failed DRY 0x1 CMD 0xef F 0x66 N 0x0 S 0x0 H 0x0 CL 0x0 CH 0x0
WARNING: /pci@0,0/pci1002,43a0@15/pci-ide@0/ide@1 (ata1):
timeout: abort request, target=1 lun=0
---end---
Any hints?Have you checked that the disc is actually readable by the CD drive which you use for booting? Perhaps it may help to use a USB stick to boot Solaris.
Kind regards,
Steffen

Dual boot Solaris 9 and Solaris 10 on a SCSI hard disk

Hi,
I would like to dual boot Solaris 9 and Solaris 10 on a SCSI hard disk. Currenlty that 16gb hard disk has Solaris 9 installed on it.
What are the steps needed to install Solaris 10 and dual boot them?
Thanks a million.
ldd

Unfortunately we can't help you as this forum is only for questions or problems relating to Sun Update Connection, Patch Manager 2.0, & PatchPro 2.2.
We recommend you post your question on one of the Solaris OS forums:
http://forum.sun.com/jive/category.jspa?categoryID=139

Booting solaris 9: /etc/bootrc?

I have installed solaris in PIII 500 and all works perfect during instalation, but when reboot to run this message appear:
<i>boot:source: open of '/etc/bootrc' failed</i>
I try to boot with the software CD1 of 2 and it say the same
I'm changing from windows and my knowledge of unix is limited, so can you give me solutions with simple words?
Thanks to all

Do you have more than one HDD installed in the system, and did you
try to install Solaris to the secondardy disk?
It seems the <font face="Courier">boot: source: open of '/etc/bootrc' failed</font>
problem is a bug in the Solaris installer. It writes the wrong bootpath
property to the root filesystem of the new Solaris installation.
<div class="pre"><pre> ----------------------</pre></div>
I just tried to install Solaris 10 b69 on c1d0s6.
c0d0 is the boot disk, master device on the primary ide channel.
c1d0 is another ide disk, master device on the secondary ide channel.
The physical device path for c1d0s6 is
<div class="pre"><pre>
# ls -l /dev/dsk/c1d0s6
lrwxrwxrwx 1 root root 50 Jan 9 2004 /dev/dsk/c1d0s6 ->
../../devices/pci@0,0/pci-ide@7,1/ide@1/cmdk@0,0:g
</pre></div>
But the installer writes this to /boot/solaris/bootenv.rc:
<div class="pre"><pre>
# grep bootpath /sol10/boot/solaris/bootenv.rc
setprop bootpath /pci@0,0/pci-ide@7,1/ide@0/cmdk@0,0:g
</pre></div>
(Note: "<a href="mailto:ide@0" target="_blank">ide@0</a>" instead of "<a href="mailto:ide@1" target="_blank">ide@1</a>")
Solaris 10 cannot be booted from the secondary disk, the Solaris boot code
reports:
<div class="pre"><pre>
Run Error: File not found
boot: source: open of '/etc/bootrc' failed
</pre></div>
Changing the bogus bootpath in bootenv.rc fixes the boot problem.
It should be possible to fix the bootpath by interrupting the boot process
with ESC, and setting the correct boot device in the "Boot Tasks" submenu
on the last screen named "Boot Solaris".

??? - Install CD w/ custom /boot/solaris/devicedb/master ?

Howdy Friends:
Any suggestion on how I might go about creating a custom Solaris 10 install CD with a modified /boot/solaris/devicedb/master file? My IDE controller isn't recognized by Solaris 10, and I need to modify the master file to include the driver info (Sil 0680 IDE RAID) for installation.
I am aware I could just create an install floppy via DD, however this system does not have a floppy drive nor a working FDD controller. I also cannot boot via a USB floppy due to no support in the BIOS.
I've tried installing via a network install (PXE), but I never get past the "Solaris Network Boot..." screen.
Any help installing on this system would be awesome. I'm excited to use Solaris 10!
-- JWheeler
tpawiredboy*at*unequa-tsi.net

The process for creating a bootable Solaris x86 CD is quite a bit harder than it is on Sparc hardware. Or at least it was with the only process I know about. It is much simpler to perform a network-based installation.
One catch to the network install, however, is that a Sparc system cannot read all the partitions on the x86 Solaris CD. I had to put the disc into an existing Solaris x86 system, NFS mount a directory from my Sparc server, and perform a "./setup_install_server" from the x86 system to the NFS directory.
Once that was complete, I could use the Sparc server to jumpstart the x86 systems using custom configurations. It also allowed me to install the correct drivers for the x86 system in question that were not part of the Solaris CD (which is one of the problems you describe).
I realize that you were probably looking for a simpler answer that this...

Installed flash disk does not boot (Solaris 10 on Fore CPU-50 VME cpu-card)

Greetings,
After installation of Solaris on a system with a flask disk as system disk, the disk is not bootable.
Configuration:
System: Fore CPU-50 VME cpu-card (sparc50)
OS: Solaris 10 (Sparc version, CD-media)
System-disk: ADTRON S35FA-8GC20N Smart Storage 3.5 SCSI Flash pack.
          (configured as c0t1d0)
Error message at boot:
Cant read disk label
Cant open disk label package
When I mount the installed disk on another system I can read and inspect the flash disk.
The logging files begin.log, finish.log and install_log in the /var/sadm/system/logs/ directory dont mention any errors during installation phase.
The logging file /var/sadm/system/logs/install_log begins with:
Configuring disk
-     Creating Solaris disk label (VTOC)
Creating and checking UFS filesystems
-     Creating / (c0t1d0s0)
-     etc.
This install_log file ends with:
Installing boot information
-     Installing boot blocks (c0t1d0s0)
-     Updating system firmware for automatic rebooting
How can I make this flash disk bootable?
Thanks,
Herm

It's a G3 with the integrated 5 i Smart array
controller. There are driver updates for Sol 10 on
the HP side. Just filter for OS. There you get the 5i
driver !
greets
D@veI searched the HP site, but wasn't able to find the driver for Solaris 10. Could you give me the url? I did find links for drivers for Solaris 8 & 9 though.
Thanks,
-Jeff

How do I install dual-boot Solaris 8 and Solaris 9 on one hard disk ?

I tried to install Solaris 8 and Solaris 9 on same disk using CDs, but
the second installation overwrote the first Solaris which was installed
previoudly on the half-disk size partition of same disk.
How do I install two Solarises on one hard disk ?
Thanks
Yakov

There are no tricks to get Solaris to dual boot on the same drive. Just allocate and pick the free slices not used by the first Solaris install when you put in the second install. Technically speaking there is nothing preventing you from running seven separately bootable Solaris instances on the same drive (one of 8 available slices is overlap -- slice 2) provided you use a swap file on a root partition instead of reserving a whole slice for swap.

Dual Boot - Solaris 8 intel under system commander

I am interesting in installing Solaris 8 for intel on my PC but need to keep my existing operating environments (98 and NT). I would like to install solaris on either a spare partition or spare disk drive.
Has anyone managed to configure solaris 8 to boot from system commander deluxe. I know it is possible with Solaris 7 but I have no documentation on for Solaris 8. I cant seem to get through to the System Commander people.

hi phil_xxx
I am using a dell dimension 4100 machine. I already have win200,win98 in the primary drive(10GB) using using 'System Commander' as boot manager to select the OS( to boot into).
I installed slave disk(30GB-IBM)and have been trying with all the possibilities to get the solaris(version-6/00)installed on the slave disk. I used "Solaris 8 Software 1/2" CD. I was not able to get solaris installed on second(slave) disk evenwitth the single user mode.
Looking for ur help badly if u have got any result.
-Nalla ([email protected])

Network Booting Solaris 10 on Netra T4 - scratch memory overflow error

Hi,
I recently got a hold of a Netra T4 server, and am now trying to install Solaris 10 on it. Since the server can DVD-driveless, I had to resort to booting it over the network from a boot server. I have progressed through a series of errors, from not configuring the NFS correctly, to making sure the RARP was on as well.
The current issue that I am having is when i start the boot (boot net -v install), it begins to load the image, and then haults when it reaches:
Rebooting with command: boot net -v install
Boot device: /pci@8,700000/network@5,1: File and args: -v install
38e00 Using RARP/BOOTPARAMS...
Internet address is: 10.10.1.201
hostname: bfc
domainname: (none)
Found 10.10.1.120 @ 0:30:48:8d:c2:6
root server: owl (10.10.1.120)
root directory: /pub/install/Solaris_10/Tools/Boot
boot: cannot open kernel/sparcv9/unix
Enter filename [kernel/sparcv9/unix]: /pub/install/Solaris_10/Tools/Boot/platfor
m/SUNW,Netra-T4/kernel/drv/sparcv9/lombus
boot: failed to allocate 8192 bytes from scratch memory
panic - boot: boot: scratch memory overflow.
Program terminated
I'm not exactly sure why this is occuring, or what I can do to resolve this. Has anyone else come across this scenario? Judging from google searches, I haven't found anything that can help me fix this. Hopefully someone can.
Thanks a ton,
Kyle

as well, I have tried using various directories to install from, all coming to the same result.

Boot Solaris 10 Install CD with Sunblade 100 with SPARCIIe

OK > boot cdrom
This fails and is looking at [platform/sun4u/kernel/sparcv9/unix], asking for the boot file name. I have used these install CDs on other systems (probably 32-bit) and I understand that 32-bit and 64-bit installs are now on same CDs.
I also tried �boot cdrom �install� with errors.
I think this error is probably related to 64-bit boot since Sunblade 100 with SPARCIIe requires 64-boot only boot.
Any idea what the boot-file I should specify is?
Thank You

The symptoms sound as if you're attempting to use CDs for the
Solaris-10(x86) instead of the proper Solaris-10(SPARC) disks.
Both sets would be readable by an OS, but only the SPARC distro
would be bootable on a SPARC-based system such as your Sunblade 100.
There hasn't been a 32-bit SPARC system since pre-Ultra days.
... perhaps 10 years ???
The PeeCee universe still swims in 32-bit-land.
The x86 port of Solaris has both kernels,
but that's for a completely different architecture.

Can't boot Solaris-10 on NetraX1

I'm trying to upgrade a NetraX1 1GB Ram (Sun Memory) 2x40GB disks over the network. The system is at Solaris-8 and runs it quite well. Trying to boot 10-u8 or even 10-u7 results in the following messages:
Sun Netra X1 (UltraSPARC-IIe 500MHz), No Keyboard
OpenBoot 4.0, 1024 MB memory installed, Serial #51330455.
Ethernet address 0:3:ba:f:3d:97, Host ID: 830f3d97.
Executing last command: boot net
Boot device: /pci@1f,0/ethernet@c File and args:
3a000 do_relocate: bad strndx 0
do_relocations: /platform/sun4u/kernel/cpu/sparcv9/SUNW,UltraSPARC-IIe do_relocate failed
krtld: error during initial load/link phase
Fast Data Access MMU Miss
ok .version
Firmware CORE Release 1.0.18 created 2002/5/23 18:22 << Upgraded using patch 111952-03 >>
Release 4.0 Version 18 created 2002/05/23 18:22
cPOST version 1.0.18 created 2002/5/23
CORE 1.0.18 2002/05/23 18:22
I set OBP diags-switch? true with diag-level set to max but no problems are reported.
Sun says the NetraX1 supports Solaris-10 but are there any caveats. It seems strange that the latest version of OBP dates to 2002. Am I trying to 'push a string' by upgrading this machine or am I doing something stupid? Any help/advice would be much appreciated. Note that the Solaris-10-u8 imaged is one I downloaded from Sun and made my own DVD but the 10/08 distribution is genuine Sun media, they both fail with the same messages.

I've tried a few other things like creating the user account in lom and most significantly tried hooking up a DVD reader in place of the second disk. The OBP must not have the driver for reading CD/DVD's as it simply wouldn't recognize the drive although I can if I boot the system from the disk under Solaris-8 and mount the CD/DVD drive.
Is there any possibility that the obp on the machine just can't handle Solaris-10?
I tried booting over the net with a mounted DVD on the host machine and using the -v switch and got the following. Can anyone interpret what's happening here? I've edited out the redundant stuff. The DVD I'm trying to boot from is a genuine Sun made DVD and not a downloaded image.
boot net -v
Res
LOM event: +0h31m3s host reset
etting ...
Sun Netra X1 (UltraSPARC-IIe 500MHz), No Keyboard
OpenBoot 4.0, 1024 MB memory installed, Serial #51330455.
Ethernet address 0:3:ba:f:3d:97, Host ID: 830f3d97.
Initializing Memory - (many times)
Executing last command: boot net -v
Boot device: /pci@1f,0/ethernet@c File and args: -v
Using RARP/BOOTPARAMS...
Internet address is: 192.168.1.94
hostname: netrax1
domainname: TheJacks
Found 192.168.1.34 @ 8:0:20:ff:fd:92
root server: ultra10 (192.168.1.34)
root directory: /cdrom/sol_10_1008_sparc/Solaris_10/Tools/Boot
/-module /platform/sun4u/kernel/sparcv9/unix: text at [0x1000000, 0x109fded] data at 0x1800000
module /platform/sun4u/kernel/sparcv9/genunix: text at [0x109fdf0, 0x125f027] data at 0x1869ac0
module /platform/SUNW,UltraAX-i2/kernel/misc/sparcv9/platmod: text at [0x125f028, 0x125f297] data at 0x18c0580
module /platform/sun4u/kernel/cpu/sparcv9/SUNW,UltraSPARC-IIe: text at [0x125f2c0, 0x126c7d7] data at 0x18c0cc0
Evaluating: 70000000000 constant kmem64-base h# 700012ba000 constant kmem64-end h# ffffffffffc00000 constant kmem64-pagemask h# e0000
000618016b6 constant kmem64-template : kmem64-tte ( addr cnum -- false | tte-data true ) if
( addr ) drop false exit then ( false ) dup kmem64-base kmem64-end within if ( addr ) kmem64-pag
emask and ( addr' ) kmem64-base - ( addr' ) kmem64-template +
( tte ) true ( tte true ) else ( addr ) pgmap@
( tte ) dup 0< if true else drop false then ( tte true | false ) then
( tte true | false ) ; ' kmem64-tte is va>tte-data
kmem64-template ?
Error in Fcode execution !!!
Evaluating: to load-base init-program
Undefined word encountered
Fast Data Access MMU Miss
ok

Booting Solaris on OpenSPARC

Similar Messages

Maybe you are looking for