hks@funet.fi (Harri Salminen) (03/16/90)
I'm not sure what's the most optimal format specification you can make for
the 1588-15 but I have at least a working one. SCSI blocks don't map
easily to sectors, heads and cylinders since they are used internally by
the drive too.
Here's my entry:
disk_type = "Micropolis 1588-15" \
: ctlr = MD21 \
: ncyl = 1625 : acyl = 1 : pcyl = 1632 : nhead = 15 : nsect = 53 \
: rpm = 3600 : bpt = 31250
partition = "Micropolis 1588-15" \
: disk = "Micropolis 1588-15" : ctlr = MD21 \
: a = 0 , 1291875 : c = 0 , 1291875
partition = "Micropolis 1588-15, large root" \
: disk = "Micropolis 1588-15" : ctlr = MD21 \
: a = 0, 89040 : b = 112, 82680 : c = 0 , 1291875 : g = 216, 1120155
With Ciprico controller I used also 1609 cylinders 54 sectors which give{~r
little bit more space and it seemed to work as well. You don't really need
those alternate cylinders either for the original defect list since the
SCSI controller does the mapping using it's own reserved sectors and
tracks. newfs has reportedly sometimes failed to give a raw device to
mkfs but I haven't had problems with that. Also installboot seems to fail
on some drives of same type if you try to use it on other than your
current root partition but you can always boot the bootloader from the
network first time... Also if you power disks before the SUN and haven't
prevented termpwr signal from loading your drive it might start to blink
it's led and seem broken although it really is O.K. When you get these
drives to work they seem to be working well and fast. They have quite an
impressive modern controller on those disks.
Hope this helps
Harri