>Has *anyone* been able to load Solaris 2.4 x86 onto a Conner 1060S 1 GB
I don't have that combination handy but I do have other 1 GB drives and
other HBAs and I haven't had any problems with those combinations with
Solaris 2.4 x86. But there have been problems reported with some brands
of 1 GB drives and some brands of controllers but as far as I know there
aren't any inherent problems with your combination. Try calling Adaptec
and Conner support and asking each of them if there any known hardware
problems with your combination.
Quote:>The setup is an AST 486/33 Bravo with 16 MB, an Adaptec 1542CF and the
>Conner with the machines BIOS turned off, the Adpatec card reset to defaults,
I don't understand what you mean by "the machine's BIOS turned off".
Quote:>(Extended BIOS translation is disabled as is Dynamic scan SCSI BUS, as
>well as BIOS support for more than 2 drives), except for DMA channel set to
>6 as stated in the Solaris guide.
Did you notice that on page 45 of the Device Configuration Guide it
also says to:
* Access the 154xC/CF BIOS setup utility by pressing the ctrl+A
hot key when prompted. Then:
* Disable Synchronous Negotiation for CD-ROM drives for
* Disable the option that supports greater than 1 GB
(1024 cylinders) on the disk (Extended BIOS
Translation for DOS Drives > 1 Gbyte).
* Disable the option that supports more than two DOS
* The DMA transfer rate for the 1540CF [this should say 154xCF]
should be left at the default unless your motherboard supports
I'm just guessing but perhaps you enabled the extended translation or
you up-ed the DMA transfer rate. Try setting your exactly as it says in
the Device Config Guide.
Quote:>Now the problem. After the install is completed, Solaris reboots and upon
>coming up it complains about /etc/init.d. Looking at the directory with
>ls -a comes back with ". not found". At this point if I reboot /etc
>gets more corrupt until on the third reboot fsck fails and the ufs
>filesystem is shot. Running fsck manually wipes out all of /etc until
>the sytem is useless.
That's very strange. Can I assume you haven't done something as simple
as forgetting to shutdown cleanly before hitting the reset button? Can
I also assume that you're not getting any SCSI I/O error messages on the
console? (Since init.d is corrupted you may not be running the syslog
daemon. I'm not certain where the messages go if syslog isn't running.
Try running dmesg after your system reboots and look for SCSI error
messages in its output.)
Quote:>This happens like clockwork each of the 3 times I have attempted
>to install Solaris.
Next time you try the install select the option which says "don't
reboot". When the install finishes run fsck -n on your freshly
installed root filesystem. If you don't have any other controllers or
disks you should do something like this:
fsck -n -F ufs /dev/rdsk/c0t0d0s0
This will at least tell us whether the filesystem is being corrupted
during the installation or later.
Quote:>I have done this using update4 disks as well as the original.
It shouldn't matter. As far as I know the aha driver hasn't been update
>The system comes up and its a running system, only the /etc/init.d
>directory is corrupt.
>I used the Adpatec 1542CF card to Low Level format the drive.
>Now what??? The drive is 2 days old and I'm not sure what to do.
Assuming you have the Adaptec HBA configured properly the other possibility
is that you're getting I/O errors but for some strange reason you're
not seeing any messages on the console.
Some early 154xCF boards were very sensitive to non-compliant SCSI
cables. I've never seen one of those boards but hopefully such a
situation will cause error messages on your console rather than silent
data corruption. Perhaps Adaptec's tech support people can tell from
your board's serial number whether you've got one of the sensitive
ones. Are your SCSI devices cabled internally or externally? If
they're internal, do you have at least a foot between connectors and
your bus correctly terminated? If they're external, are you using good
quality cables (I know Adaptec sells compliant cables)?
You didn't say what jumpers you've set on the disk drive. I don't know
specifically about the Conner 1060S, but most drives have jumpers for
parity enable and SCSI-2 mode. Both of these options should always be
If none of this helps contact Solaris x86 support and they can dig up
the exact same hardware and verify that there isn't a hidden bug.