I disagree ... based on extensive experience with Solaris per se. But I
realize that zfs is a different animal from ufs so I just looked at my
Solaris 10 ultra24 /etc/vfstab file to see where the root filesystem
mounts. And sure enough, that's no longer in /etc/vfstab with a zfs
root filesystem.
OK, I best defer to others for more detail but you can get some insight
by executing 'strings -a /etc/zfs/zpool.cache' and looking at the device
related content (in my case I see /dev/dsk/c1d0s0, cmdk@ASEAGATE...,
pci@0,0/pci-ide@1f,2/ide@1/cmdk@0,0:a, whole_disk among other entries).
You will not see this until you correct the discrepancy, I expect,
because your kernel hasn't found the existing disk with the
corresponding cache yet, even though it booted from it!
What has happened in your case is something along the lines of -
You were plugged into cmdk@0,0 when you installed.
you are now plugged into cmdk@1,0 now.
The boot process loaded the kernel without utilizing zfs.
It now cannot mount the root filesystem because the disk is at the wrong
target number.
If you get the SATA cable back in the matching motherboard port I don't
think you have to do anything else. OTOH, if you need to leave it in
the non-matching port, start by using a naked 'zpool import' to see if
that will list your existing pool, if you see it you can rerun the
import with the additional arguments needed to actually import it (or
'zpool import -a' will import everything it finds).
On 03/14/11 15:12, Roy Sigurd Karlsbakk wrote:
Not really - Solaris doesn't rely on where a drive is placed - as he said, the
OS comes up correctly, but some services might not start.
----- Original Message -----
Not right.
<snip>
--
Jerry Sutton jer...@airmail.net
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
