OK problem solved.
I had incorrectly assumed that the server wasn''t booting, the longest I
had left it was over night and there was still no logon prompt in the morning!
The reality is that it was just taking a very long time due to an excessive
amount of automatically created snapshots by Time Slider.
It did occur to me earlier that the amount of snapshots might cause delays
mounting the pool as I had previously left Time Slider on for all filesystems
and volumes. Back then I used Time Slider to erase the snapshots, again,
incorrectly assuming it would also erase the volume snapshots, I can now confirm
that it does not!
So, to resolve, I reinstalled the OS (installing another instance to another
disk would work), imported the array, the import command then seems to get stuck
(obviously, in hindsight), although the core filesystems are present and
accessible. Now, before rebooting, I then run this command to erase all
automatically created snapshots:
zfs list -r -t snapshot -o name -H array/iscsi | grep zfs-auto-snap | xargs -t
-l zfs destroy
Obviously, this takes while ;-)
After this completes, I reboot and all is well again!
--
This message posted from opensolaris.org