Creating Custom Solaris 9 x86 boot CDs

Does anyone know how to create custome Solaris 9 boot CDs for the x86 platform? I want to try and remove some of the packages that I don't need and see if I can get the boot CDs down to 1 CD (as I did with the Solaris Sparc version). I tried the steps layed out in "Building a bootable JumpStart Installation CD-ROM", but the CDs do not appear to be in the correct format. If I put these new CDs into a SPARC station, they appear to be layed out correctly. After some research I determined that the x86 CDs use El Torito with the .boot-image in the s2 slice, but it still doesn't work.
Here are the steps I have tried:
From the Jumpstart whitepaper
1. Extract slice 2 from the CD: find . -print | cpio -pudm /bicd/s2
2. Stop volume management: /etc/init.d/volmgt stop
3. Extract slice 0 from the CD: dd if=/dev/dsk/c1t0d0s0 of=/bicd/s0.image
4. Modify /bicd/s2
5. Make an ISO file: mkisofs -R -d -L -l -o /bicd/s9.image -B /bicd/s0.image /bicd/s2
6. Burn the CD: cdrw -i s9.image
I modified step 5 for an El Torito image:
Make an ISO file: mkisofs -R -d -L -l -o /bicd/s9.image -B /bicd/s0.image -b .boot-image /bicd/s2
The Jumstart whitepaper says to extract the vtoc separate, then calculate the cylinder boundaries and include it on the new CD. Whenever I include the extracted vtoc, the CD is completely unreadable. If I leave this step out of the Solaris Sparc version, everything works okay.
Any help would be appreciated. Thanks.

Similar Messages

Solaris 10 x86 Boot Issue

Hi,
I've installed Solaris 10 for x86 into my P4 machine. It has three partitions - two of them are being used by Win2K Advanced Server and in the 3rd one Solaris 10 was installed. Installation was successful and the Boot manager is working fine. I can boot windows from the Boot Manager. But Solaris is not booting. It is freezing afer I give enter in proceed with Default booting. The 'SunOS.....' Message is coming and that's it. Even no Erro message came also.
Need some help.
Thanks

Solaris makes one of the partitions active , Take a note of that.
Once you are done tinkering with the partitions/ loading say Windows/Linux follow this step.
Grab any Linux distro's boot CD and start the rescue mode. type grub and do this
root (hd0, X) X the partiton that has linux partition minus 1
setup (hd0)now open fdisk and change the label of the Solaris Partition to 80.
Add an entry in /boot/grub/menu.1st on the same lines like Windows, If unsure google .
Reboot and you are done.
Never I said never make any other partition active other than what Solaris installation does for you.

Solaris 10 x86 boot from SAN

Hello
I am trying to install solaris10 x86 on an IBM Blade booting from SAN. The blade includes a fibre channel extension card (qla2312 from qlogic) and no internal drive.
The installation procedure does not find any bootable device to install solaris on. I have tried the qlogic driver disk for solaris 9, the driver is loaded (apparetly), but no disk is found (a lun is configured and presented to the blade, the bios of the FC card can see it).
Is there any solution ?
thanks in advance

I just today posted in "Solaris 10 General" about this same topic.
It appears that only the sparc supports this but I'm not certain.
As I stated in my other post and as you also state, the installer doesn't see the lun.
fyi, I was able to do this with RHEL on the same blade and to the same lun.
Did you find any solution ?

Solaris 10 x86 boot menu configuration

Is there a way to change its configuration? ...Currently I have WindowsXP installed on partition 0 and Sol 10 on partition 1. The boot menu lets me select between the Windows partition and the Solaris partitions but it defaults to the S10 one. I would like to boot Windows every time unless I specifically tell it otherwise but if I mark the Win partition active, I lose the boot menu - the S10 partition needs to be active.
I tried copying mboot to the root of the Windows partition adding an entry to boot.ini but says: "No active partitions" when I select it at boot time.
I saw that Sol 11 uses Grub. Can I use it with Sol 10 also? If so I'd appreciate help setting it up.
What are my options?
TIA :)

Thanks to your reply.
To be on the safe side, I tried to install Sol 10 on a used hard disk of 20GB that was lying around. Don't want to mess with the other hard disk of 120GB which contains lots of valuable info.
The 20GB hard disk had some bad sectors in the past. Booted the system from CD ROM and the OS started to collect hardware info. Glad it detected video card, monitor,key board and mouse. Proceeded with the installation after completing the intial phase of selecting the country,time and others selections.
After that, I ran into a problem with the following error repeating continously. Had to shutdown the system.
PCI@0,0/PCI-ide@4,1/ide@0/cmdk@0,0 (Disk0)
Error for command 'read sector" error level - retryable
Required Block - 22177 Error Block - 22177
Sense Key: bad block detected
Vendor 'Gen-ATA" Error Code: 0x8
I will take your suggestion of creating another primary partition on the current hard disk of 120GB with a partition size of 20GB for Sol 10. Will keep the forum posted.
Ram

Solaris 10 x86: Booting from floppy to install

Hi all, I have an old PIII machine which won't boot from CD.
Does anyone know how to generate the floppyes that could boot and then access the CD like in Solaris 2.6 ?
Thank you very much!
M

What revision of Solaris 10? I had this happening with an older release of Solaris 10 on my desktop, and it magically fixed itself with release 8/07.

How-To create a customized Solaris 10 DVD for Sparc with GRUB

I am currently using Solaris 10 U6 x86 custom jumpstart from DVD. This boot mechanism has always been Grub. I know that Solaris 10 U8 Sparc boot mechanism has been changed to Grub as well. I have referenced an excellent Blueprint on creating customized jumpstart DVD for x86. I am looking for some documentation or Blueprint which describe the similar process for Sparc. I am hoping the mkisofs command options will be the most significant difference between x86 and Sprarc.
Is there a document which describes the de-composition of the DVD for Sparc based platforms? And the re-assembly of the custom DVD?
Thanks very much.
John

Hello Aruna,
As far as I know you cannot remove the default items for the menu.
Sophie

Boot image for Solaris 8 x86?

Does anyone know where I can download the image for the boot floppy for Solaris 8 x86?
Mine got corrupted, and since System Commander 2000 doesn't recognize the partition, that and the CD are the only ways I can boot the system.
Thanks in advance,
-Phil
pbflores@"REMOVETHIS"usa.net

Are you considering any particular cards ?
I have had good luck with DEC21040 (10baseT) and DEC21140 (100baseT) cards
in Solaris 7 through Solaris 10 on x86.
The same can be said for 3Com 3C905C-TX cards.
Solaris 8x86 will recognize many chipsets natively,
and a number of third party developers have created working driver packages.
Most of their driver packages are free, but a few are not free.
Solaris 9 works with a few more, and Solaris 10 works with yet more adapters.
The Solaris 8x86 HCL list has the native cards.
This other Sun web page can direct you toward 3rd party sites.

Create filesystems on Solaris 10 x86

Hi All,
I am not expert in Solaris, I just want to add an external
disk (via San) to a Solaris 10 x86 system.
When I execute #format, have the following :
0) c1t0d0 ....
2) c4t60060160D9E01D009AAB136E0F54DE11d0
/Scsi_uhci/disk@g60060160d9ede11
When I choose number 2, the very disk, I have the following :
Selecting c4t60060160D9E01D009AAB136E0F54DE11d0
[disk formatted]
/dev/dsk/c4t60060160D9E01D009AAB136E0F54DE11d0s2
is part of SVM volme stripped : d7
Please see metaclear.
Then I execute "verify" at the format Promp and got :
Part Tag Flag Cylinder Size
0 Unassigned wm 0 0
1 Unassigned wm 0 0
2 Backup wu 0-17442 133.62Gb
3 Unassigned wm 0 0
4 Unassigned wm 0 0
5 Unassigned wm 0 0
6 Unassigned wm 0 0
7 Unassigned wm 0 0
8 Boot wu 0 7.84Mb
9 Unassigned wm 0 0
My problem is to create 2 partitions, one of 70Gb and the
other one is the rest of the disk, then create filesystems on
them, how to proceed ?
Thanks
Bandon

Hi All,
I am not expert in Solaris, I just want to add an external
disk (via San) to a Solaris 10 x86 system.
When I execute #format, have the following :
0) c1t0d0 ....
2) c4t60060160D9E01D009AAB136E0F54DE11d0
/Scsi_uhci/disk@g60060160d9ede11
When I choose number 2, the very disk, I have the following :
Selecting c4t60060160D9E01D009AAB136E0F54DE11d0
[disk formatted]
/dev/dsk/c4t60060160D9E01D009AAB136E0F54DE11d0s2
is part of SVM volme stripped : d7
Please see metaclear.
Then I execute "verify" at the format Promp and got :
Part Tag Flag Cylinder Size
0 Unassigned wm 0 0
1 Unassigned wm 0 0
2 Backup wu 0-17442 133.62Gb
3 Unassigned wm 0 0
4 Unassigned wm 0 0
5 Unassigned wm 0 0
6 Unassigned wm 0 0
7 Unassigned wm 0 0
8 Boot wu 0 7.84Mb
9 Unassigned wm 0 0
My problem is to create 2 partitions, one of 70Gb and the
other one is the rest of the disk, then create filesystems on
them, how to proceed ?
Thanks
Bandon

Going back to mirrored after booting unmirrored - Solaris 10 x86

Folks,
On one of my Solaris 10 x86 servers at the last reboot it failed to come up. I put back the versions of /etc/vfstab and /etc/system from before the boot drive was mirrored. "metastat" still reports the md versions of the slices are still in sync but I'm not convinced that's accurate since the failed reboot.
How do I go back to mirrored again at the next boot?
I've been reading doc 819-2789 "x86: How to Create a RAID-1 Volume From the root ..." and it looks like I need to go through most of the steps.
My best guess goes something like this -
Loop for each mounted slice - Use metadetach to remove the one not mounted. Use metaattach to add it back. Use metasync to confirm the mirror synced up again.
Once that loop has completed assume that any MBR stuff is broken. A bit like I've done with "mkboot" on SPARC but following the steps in the doc. Look at menu.lst, fdisk, installgrub, metaroot, have with and without mirror copies of /etc/vfstab and /etc/system ready just in case, reboot.
Basically all of the steps from the doc only instead of using metainit use metadetach, metaattach, metasync.
Does this make sense? It parallels stuff I've done on SPARC but I have a lot more experience on SPARC than on X86.
Thanks in advance, Doug Freyburger

freyburgerd wrote:
Folks,
On one of my Solaris 10 x86 servers at the last reboot it failed to come up. I put back the versions of /etc/vfstab and /etc/system from before the boot drive was mirrored. Exactly how you did that may matter.
"metastat" still reports the md versions of the slices are still in sync but I'm not convinced that's accurate since the failed reboot.You're absolutely correct. As soon as you write (or mount read/write) to one of the raw slices or to one of the sides of the mirror, the mirror will be out of sync, and SVM will not realize it.
How do I go back to mirrored again at the next boot?Detach the side you're not booted from. Redo the metaroot and come up on the (now one-way) mirror. Once booted from the mirror, reattach the other side.
My best guess goes something like this -
Loop for each mounted slice - Use metadetach to remove the one not mounted. Use metaattach to add it back. Use metasync to confirm the mirror synced up again.You'll want to detach before using the mirror again, but you can't attach until you stop using the raw slice. So you can't just detach/attach in one step. You'll need to reboot between them.
Once that loop has completed assume that any MBR stuff is broken. A bit like I've done with "mkboot" on SPARC but following the steps in the doc. Look at menu.lst, fdisk, installgrub, metaroot, have with and without mirror copies of /etc/vfstab and /etc/system ready just in case, reboot.
Not sure why any MBR stuff would have been affected unless you did so explicitly.
Basically all of the steps from the doc only instead of using metainit use metadetach, metaattach, metasync.
Does this make sense? It parallels stuff I've done on SPARC but I have a lot more experience on SPARC than on X86.Should be the same on both. If you're adding a new disk, you need to make sure boot support is added. But in this case I doubt any of the boot support was damaged (depending on the details of what led up to your changes).
Darren

Creating Bootable Solaris CDs on Linux

Hi,
I downloaded the zipped iso images for the Solaris 8
(7/01) CDs. I tried to create the CDs using cdrecord
but the Installation CD and Software CD1 are coming
out as un-bootable CDS.
Here is a command I used:
cdrecord -v -eject speed=12 dev=0,0,0 sol-8-u5-sparc-v1.iso
Similarly for the Installation disk.
Is there something else I should do the make CDs
bootable on a Sparc machine?
The Installation CD shows only 120KB of data. It has a erroneous link (Boot -> ../s1) in the root directory.
Any help is greatly appreciated.
Thanks for your time.
Sandeep

Problems burning the cd's range from bad downloads to improper option settings in the burning software...take a look at these prior discussions and see if any of the suggestions help
http://forum.sun.com/thread.jsp?forum=11&thread=3906
http://forum.sun.com/thread.jsp?forum=11&thread=3527
http://forum.sun.com/thread.jsp?forum=11&thread=12109
http://forum.sun.com/thread.jsp?forum=11&thread=11630
I used easy cd creator without any problems although Nero is suggested a lot. I hope this can be of some help. You may also want to search the forum for posts about your cd software.
Lee

Not able to create database in Solaris 10 x86 Oracle 10.1.0.3

I have installed Oralce 10.1.0.3 in Solaris 10 X86 using the procedure given without creating the database.(Installing the binaries)
Hardware:Sun fire V20z.
When i tried creating Database using DBCA,it was throwing lots of error.
The errors are:
1) ORA-03113: end-of-file on communication channel
2) ORA-03114: not connected to oracle
3) ORA-27100: shared memory realm already exists.
Any help on creating a database in Solaris 10 x86 Oracle 10.1.0.3
Thanks in Advance.

Hi.
Are You read documentation ?
[http://download.oracle.com/docs/cd/E19680-01/html/821-1255/babhabac.html#z40000471052274]
Guidelines for the /globaldevices File System
Oracle Solaris Cluster software offers two choices of locations to host the global-devices namespace:
A lofi device
A dedicated file system on one of the local disks
This section describes the guidelines for using a dedicated partition. This information does not apply if you instead host the global-devices namespace on a lofi.
The /globaldevices file system is usually located on your root disk. However, if you use different storage on which to locate the global-devices file system, such as a Logical Volume Manager volume, it must not be part of a Solaris Volume Manager shared disk set or part of a VxVM disk group other than a root disk group. This file system is later mounted as a UFS cluster file system. Name this file system /globaldevices, which is the default name that is recognized by the scinstall(1M) commandSo:
>I have mounted one (1) common 2gb LUN for both the nodes, so it is creating the problem...?
It source of problem.
Every node require have it's own separate slice for globaldevice.
Shared Lun can be mounted with option global for access from both nodes at same time, but it will be second step - configure data-services.
Regards.

Solaris 10 x86 - software RAID boot

What I can write in eeprom : - for boot form 1-st, and if it absent - from 2-nd disk ? (2 disks in soft raid). two bootpath ?
In Solaris 9/sparc for boot with one working HDD in mirror (2 disks in mirrir) I wrote in /etc/system : set md:mirrored_root_flag=1
What I can wrote for this in solaris 10/ x86 ?

Just for the record: The problem is solved.
I went on looking for more info on this message:
Oct 14 00:07:09 schille gda: [ID 107833 kern.notice] Sense Key: ICRC error during UDMAAnd then found an interesting post from 2004 by Juergen Keil:
The "bad block detected" message appears to be an old/obsolete error
message text, somewhere in the Solaris "ata" driver.
The ATA-1 standard (1997?) defined the 0x80 bit in the IDE
controller's error register as "bad block detected". (See page 32 on
<URL:http://www.t13.org/project/d0791r4c-ATA-1.pdf>)
The 0x80 bit seems to be "reserved" (undefined, not used any more)
with ATA-2 and ATA-3.
ATA-4 and newer now defines this bit 0x80 as ICRC "interface CRC error
has occurred during an Ultra DMA data transfer" when using ultra dma
transfers. (see the description of the "READ DMA" or "WRITE DMA"
command in <URL:http://www.t13.org/project/d1153r18-ATA-ATAPI-4.pdf>).
The Solaris ata driver seems to interpret the 0x80 bit according to
the old / obsolete 1997 ATA-1 standard.
So the "bad block detected" message seems to be a hint that some sort
of data corruption is happening on the IDE interface during UDMA data
transfers with the HDD. With UDMA capable disks it apparently does
not mean "the disk drive is dying".
I went to search what could cause a hardware datacorruption on my system and found that my internal removeable disktray was making bad contact.
After replacement everything worked as expected.

Problem booting a Solaris 10 x86 after Jumpstart Installation

Hi,
I have a strange problem. I have a working x86 machine with Solaris 10 installed. I have made a flash image of it's hard disk, and I have store thar flar on another machine. I have another Solaris 10 with Jumpstart.
I boot other x86 machine through PXE. The Jumpstart machine gives it a valid IP, and the process of sending the flash image begins. The process finish, and the client machine reboots. And here is the strange problem:
When starting, the Grub men� appears, and shows the two options configured: first, start Solaris in normal mode, and second, failsafe.
If I choose normal start, the screen goes black and the machine restart as if you rebooted it. I can only start in failsafe mode.
Here are the config files in Jumpstart for that machine:
Profile:
install_type flash_install
archive_location nfs 10.95.8.7:/export/backup/platform/alba-ucm/ALBA-UCM-HostID_2b72c461_4x80GbMirror-Prepared_09042008.flar
partitioning explicit
filesys c0t0d0s0 1:2650 /
filesys c0t0d0s1 2651:1600 /var
filesys c0t0d0s3 4251:400 swap
filesys c0t0d0s4 4651:1000 /opt/app
filesys c0t0d0s5 5651:3143 /repositoryCC
filesys c0t0d0s6 8794:100 /tmp
filesys c0t0d0s7 8894:26
filesys c0t2d0s0 1:8093 /datos
filesys c0t2d0s1 8094:800 /DBoracle
filesys c0t2d0s7 8894:26
After Script:
touch /a/etc/notrouter
touch /a/noautoshutdown
sync
echo "Changing and syncing bootenv.rc"
# clear the boot-args property
echo "setprop boot-args ''" >> /a/boot/solaris/bootenv.rc
# set the bootpath property to boot from the hard disk
STRING=`df | grep '^/a '� | sed 's/).*//' | sed 's/^.* (//'`
STRING=`ls -l ${STRING}`
MYROOT=`echo $STRING | sed 's/.*..\/..\/devices//'`
echo "setprop bootpath ${MYROOT}" >> /a/boot/solaris/bootenv.rc
# disable kdmconfig from running after the first reboot
sysidconfig -b /a -r /usr/openwin/bin/kdmconfig
sync
And here is the menu.lst for Grub:
title Solaris 10
root (hd0,0,a)
kernel /platform/i86pc/multiboot
module /platform/i86pc/boot_archive
title Solaris Failsafe
kernel /boot/multiboot kernel/unix -s
module /boot/x86.miniroot-safe
Please, any idea about what may be hapening?

Thanks for your answers.
Yes, both machines have the same hardware exactly. Finally, I could find where the problem was.
The file bootenv.rc, for some reason that I don't know, was corrupted: In it there was a correct line with bootpath variable correctly defined, but, at the end of the file, there was another line setting the bootpath variable to a long and incoherent string of chars. Booting in failsafe, mounting the system in /a, and removing that stupid line from the bootenv.rc, solved the problem.

Boot from second hard drive with ZFS in Solaris 10 x86

Hi,
The usual menu.lst to boot Solaris10 x86 with a boot environment that contains a ZFS boot loader is this one
title Solaris 10 5/08 s10x_nbu6wos_nightly X86
findroot (pool_rpool,0,a)
kernel$ /platform/i86pc/multiboot -B $ZFS-BOOTFS
module /platform/i86pc/boot_archive
title Solaris failsafe
findroot (pool_rpool,0,a)
kernel /boot/multiboot kernel/unix -s    -B console=ttyb
module /boot/x86.miniroot-safeI understand that to add an alternate boot disk, I have to play with installgrub and then attach it in the pool.
But my question is : Is there a way to boot from my alternate hard disk by adding a new entry in menu.lst or is the only
way is to modify the BIOS parameter?
Thanks for your help,
Groucho_fr

Groucho_fr wrote:
Thanks for your answer Alan,
But ok, I imagineI can add a grub entry. But what I want to know is what is the exact way to boot from an alternate hard disk on Solaris 10u6
x86.AFAIK there are only two ways to do this. Select the disks via BIOS which becomes a royal PITA in a short amount of time or play with a boot manager such as GRUB until you get the syntax right.
Personally I put all OS's on the primary disks if possible to avoid these situations and if I have others to play with then you can easily mount them in or set them up as D: drive etc...
The only exception would be in an enterprise setting where you need to do things such as mirroring but then again in those situations it's only one OS per machine anyways so it's much easier.
So at the end, I am not sure what I have to configure and what is the procedure to boot from my alternate boot diskTry the grub homepage for the grub manual and hopefully for a mailing list or try a search engine.
http://docs.sun.com/app/docs/doc/819-5461/zfsboot-1?a=view
has an example to follow of using -B.
And moreover, the installgrub does not working for me :
root@mac1 # zpool status
pool: rpool
state: ONLINE
scrub: resilver completed after 0h2m with 0 errors on Thu Mar 26 10:56:24 2009
config:
NAME          STATE     READ WRITE CKSUM
rpool         ONLINE       0     0     0
mirror      ONLINE       0     0     0
c5t0d0s0 ONLINE       0     0     0
c5t4d0s0 ONLINE       0     0     0
errors: No known data errors
root@mac1 # installgrub /boot/grub/stage1 /boot/grub/stage2 /dev/rdsk/c5t4d0s0
cannot open/stat device /dev/rdsk/c5t4d0s2
You shouldn't need to do an installgrub if you installed Solaris. Try c5t4d0s0 not s2.
alan

Creating a Non-RAC ASM 10gR2 instance in Solaris 10 (x86)

DB Version:10gR2
OS: Solaris 5.10 for x86 (Installation DVD released on June 2006)
I have decided to remove the windows installation in my old laptop.I am going to install only Solaris 10 (x86) in this to create an ASM instance.
My laptop has a single IDE Harddrive. Is it possible to create ASM slices on this ?

Bjoern, Rajesh,
Thank your both very much.
Bjoern,
I have downloaded the latest solaris 10 for x86 (sol-10-u8-ga-x86-dvd.iso) from sun/oracle website . Need to write this to a DVD and test the installation.
Edited by: resistanceIsFruitful on May 13, 2010 5:50 AM

Creating Custom Solaris 9 x86 boot CDs

Similar Messages

Maybe you are looking for