partitioning RD54
Eric Smith
I've got an RD54 on an RQDX3, and I'm trying to partition it and create
filesystems. I did a "disklabel -w ra0 rd54". I can then do a newfs
on /dev/rra0a, /dev/rra0d, and /dev/rra0e, but I get an error when I try
to newfs /dev/rra0f:
# newfs /dev/rra0f
ra0: invalid command (unknown subcode) (code 1, subcode 224)
write error: 190484
wtfs: Undefined error: 0
"disklabel ra0" produces (before any newfs):
# disklabel ra0
# ra0c:
type: MSCP
disk: rd54
label:
flags:
bytes/sector: 512
sectors/track: 17
tracks/cylinder: 15
sectors/cylinder: 255
cylinders: 1225
total sectors: 312375
rpm: 3600
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0 # milliseconds
track-to-track seek: 0 # milliseconds
drivedata: 0
7 partitions:
# size offset fstype [fsize bsize cpg]
a: 15884 0 unused 1024 8192 # (Cyl. 0 - 62*)
b: 33440 15884 swap # (Cyl. 62*- 193*)
c: 311200 0 unused 1024 8192 # (Cyl. 0 - 1220*)
d: 15884 49324 unused 512 4096 # (Cyl. 193*- 255*)
e: 55936 65208 unused 1024 8192 # (Cyl. 255*- 475*)
f: 190485 121144 unused 1024 8192 # (Cyl. 475*- 1222*)
g: 261876 49324 unused 1024 8192 # (Cyl. 193*- 1220*)
I'm sure I must be doing something exceptionally stoopid, but I'm not sure
what.
Is there any particular reason why the disktab entry specifies 1225 cylinders,
rather than the manufacturer-specified 1224? The end of partition f looks
like it would be at sector 311628, somewhat lower than the 312120 sectors
the drive should have or the 312375 sectors given in the geometry section.
How many sectors does the RQDX3 actually make available on an RD54?
I'm hoping to get an KFQSA and RF35 working soon, but for now the RD54 is
the only suitable disk drive I've got.
Thanks!
Eric
POR...@trailing-edge.com
>Is there any particular reason why the disktab entry specifies 1225 cylinders,
>rather than the manufacturer-specified 1224?
Because the folks who constructed the disktab entry didn't know what
what they were doing.
> The end of partition f looks
>like it would be at sector 311628, somewhat lower than the 312120 sectors
>the drive should have or the 312375 sectors given in the geometry section.
>How many sectors does the RQDX3 actually make available on an RD54?
Gack!!! I ranted and raved about this here 3 or 4 years ago, and they *still*
haven't fixed this brain-damaged disktab entry? Well, time to rant and
rave again:
A RD54 on a RQDX3 has 311300 "normal" blocks. Those "other" sectors are
reserved for the RCT and for replacement blocks. The numbers calculated
from the drive's physical geometry are meaningless. They have been
meaningless since the first UDA50 rolled off the assembly line almost
a quarter-century ago, giving us freedom from having to know the physical
geometry of a drive.
Well, that rant and rave probably had as much effect (ZERO) as the last
time I did it, but don't say I didn't warn you. In particular, on many
MSCP disk controllers if you issue writes past the end of the "user
area" of the disk you'll clobber the RCT. If you do that, well, you're
almost totally screwed. If you've got XXDP+ on a PDP-11 you can use
ZRQC?? to go in and salvage the RCT in some circumstances, but on a
MicroVAX you're screwed because the field service diagnostics can't
salvage the RCT. Heck, the field-service diagnostics are barely capable
of formatting an already good RD54 into a RD54.
As to why you get a write error trying to newfs your RD54, I don't think
the brain-damaged partition table is the reason here, as you almost certainly
aren't running up against the end of the disk.
Tim.
David Brownlee
> Gack!!! I ranted and raved about this here 3 or 4 years ago, and they
> *still* haven't fixed this brain-damaged disktab entry? Well, time to
> rant and rave again:
>
Do you have the open PR number for the problem? - I'll see if I
can get it committed.
David/absolute