> I am trying to install nv 35 to a Dell Latitude LS which has 256MB memory and
> a Fujitsu HK2960AT hard drive.
> The install runs fine up to the point where it try's to build the file system
> on the root device which fails.
> Creating and checking UFS file systems
> - Creating / (c0d0s0)
> system creation failed for / (c0d0s0)
>
> ERROR: Could not check or create system critical file systems
Hmm, a Solaris x86 ata driver problem?
> then ran newfs /dev/rdsk/c0d0s0 which completed after a long pause and
> several warnings:-
> # newfs /dev/rdsk/c0d0s0
> newfs: constructing a new file system
> /dev/rdsk/c0d0s0: (y/n) y
> /dev/rdsk/c0d0s0: 10657792 sectors in
> 2602 cylinders of 128 tracks, 32 sectors
> 5204.0MB in 101 cyl groups (26 c/g,
> 52.00MB/g, 6400 i/g)
> uper-block backups (for fsck -F ufs -o b=#) at:
> 32,106560, 213088,319616 ...........
>
> WARNING: /pci at 0,0/pci-ide at 7,1/ide at 0/cmdk at 0,0 (Disk0):
> Error for command 'write sector'
> Error Level: Informational
> Sense Key : aborted command
> Vendor 'Gen-ATA ' error code: 0x3
...
This seems to confirm that there is a problem with the ata disk and / or the
Solaris ata disk driver.
> Any ideas how I can install Solaris to this system?
Do you know what kind of IDE chipset is used in that Dell Latitude LS system?
A few days ago I've observed a similar problem when using a Promise PDC20265
PCI IDE bus master controller. This is filed as bug 6414475.
For the Promise IDE controller, one workaround was to disable DMA data transfers
for the ATA disks, for example by defining an eeprom property "eeprom
ata-dma-enabled=0";
or by adding the following line to /platform/i86pc/kernel/drv/ata.conf:
pci-ide-blacklist=0x105a,0xffff,0xd30,0xffff,2;
(the above pci-ide-blacklist entry only works for the Promise PDC20265
controller).
At installation time you can try to experiment with the "ata-dma-enabled=0"
property.
At the GRUB boot prompt, type 'e' twice to edit the kernel boot commands, and
change
the boot command line so that
-B install_media=cdrom,ata-dma-enabled=0
is passed to the kernel. 'ESC' and 'b' boots with the changed command line.
Does that help?
This message posted from opensolaris.org
