There are known cases where the use of 'boot_disk'
isn't able to resolve to a device on x86. See the following
Heads Up which introduced support for 'boot_disk':
http://hub.opensolaris.org/bin/view/Project+caiman/Heads+upFlag+day+for+bugs+5451%2C7057%2C7058
Namely, the following message implies there are some
BIOSes that can support this and some that can't (though
the bug referenced doesn't really characterize which
BIOSes can and can't.)
* some x86 BIOSes don't report boot disk correctly -
see following bug for more details:
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6831009
At any rate, if you're seeing this behavior, the simple
alternative is just to specify the explicit disk name (ctds)
in your manifest.
-ethan
On 04/12/10 10:21, Reed Liu wrote:
Hi Shawn,
After applying the workaround, no complain about "bad geometry" any
more. But it will still fail.
...
r...@opensolaris:~# cat /tmp/install_log
<AI Apr 12 10:14:54> /tmp/ai_combined_manifest.xml is a valid manifest
<AI Apr 12 10:14:54> Auto reboot enabled
<AI Apr 12 10:14:54> Initiating Target Discovery...
<AI Apr 12 10:14:56> Target Discovery finished successfully
<AI Apr 12 10:14:56> Boot disk specified as installation target, but
the boot disk was not found
<AI Apr 12 10:14:56> ai target device not found
r...@opensolaris:~# format
Searching for disks...done
AVAILABLE DISK SELECTIONS:
0. c4t4d0 <DEFAULT cyl 8921 alt 2 hd 255 sec 63>
/p...@0,0/pci8086,2...@4/pci1000,3...@0/s...@4,0
1. c4t5d0 <DEFAULT cyl 17845 alt 2 hd 255 sec 63>
/p...@0,0/pci8086,2...@4/pci1000,3...@0/s...@5,0
Specify disk (enter its number):
r...@opensolaris:~# uname -a
SunOS opensolaris 5.11 snv_135 i86pc i386 i86pc
...
I'd like to know if someone else has ever encountered the same
problem, and how to fix it.
thanks,
-Reed
Shawn Walker wrote:
On 04/12/10 12:56 AM, Reed Liu wrote:
Hi Folks,
My installation always failed due to bad geometry.
.......................
r...@opensolaris:~# uname -a
SunOS opensolaris 5.11 snv_135 i86pc i386 i86pc
r...@opensolaris:~# cat /tmp/install_log
<AI Apr 11 21:14:20> /tmp/ai_combined_manifest.xml is a valid manifest
<AI Apr 11 21:14:20> Auto reboot enabled
<AI Apr 11 21:14:20> Initiating Target Discovery...
<OM Apr 11 21:14:21> Ignoring c4t4d0 because of bad Geometry
<OM Apr 11 21:14:21> Ignoring c4t5d0 because of bad Geometry
<AI Apr 11 21:14:22> No disks found on the target system
<AI Apr 11 21:14:22> Automated installation failed in Target Discovery
module
<AI Apr 11 21:14:22> Please see previous messages for more details
...
I can read/write/format my disks without any errors. How can avoid the
problem of bad geometry?
This might be:
15294 installer says "no disks were found" but format / gparted
sees them
http://defect.opensolaris.org/bz/show_bug.cgi?id=15294
Tracked in bugster as:
DM_NHEADS and DM_NSECTORS not available for disks w/o VTOC starting
build 135
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6937739
Workaround:
Before the installer is launched, manually create Solaris partition
on target disk by means of Solaris fdisk(1M). That triggers creating
default VTOC as well.
Cheers,
-Shawn
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
