mcm@rti.rti.org (Mike Mitchell) (08/11/90)
We recently bought an RZ57 disk drive for our DECstation 3100. We are
running ULTRIX 3.1A on the DECstation, so we were missing an RZ57 entry
from our "/etc/disktab" file. We contacted another site that had
ULTRIX 4.0, and this is what they sent us:
rz57|RZ57|DEC RZ57 Winchester:\
:ty=winchester:ns#71:nt#15:nc#1925:\
:pa#32768:ba#8192:fa#1024:\
:pb#184320:bb#4096:fb#1024:\
:pc#2025788:bc#8192:fc#1024:\
:pd#299008:bd#8192:fd#1024:\
:pe#299008:be#8192:fe#1024:\
:pf#596284:bf#8192:ff#1024:\
:pg#614400:bg#8192:fg#1024:\
:ph#1194300:bh#8192:fh#1024:
Now then, the manual that came with the BA42 box says that an RZ57 has
1,954,050 usable disk blocks on it, and that is what 'chpt' says before
I muck with the partition table. The disktab entry says the 'C' partition
should be 2,025,788 blocks! Even after mucking with the table, I cannot
read past block 1,954,049. Not only that, but 71 sectors/track * 15
tracks/cylinder * 1925 cylinders equals 2,050,125 NOT 2,025,788!
OK, so now I know that the disktab entry is wrong. The number of cylinders
must be wrong, so I look in the manual. There is nothing there (directly)
about the number of cylinders. It does imply there are 15 tracks/cylinder
(1.0GB / 66.7MB per surface), and it does say there are 71 blocks / track.
But it also says there are 35,976 bytes per track, which is only 70.27 blocks
per track. Something here is just not right. Maybe they are including
the spare sectors for bad-blocks! Lets see, it says there are 11 spare sectors
per cylinder, so there should be (71 - 11/15) = 70.27 blocks per cylinder.
Well, thats just dandy! Remember, the disktab entry should contain the
number of usable sectors on the disk, and the spare blocks are not directly
usable. What should I set the number of sectors/track too? I guess 70,
because 70 is a lot closer to 70.27 than 71. Hmmmmm... 1,954,050 is divisible
evenly by 70, but not by 71. Not only that, but 1,954,050 is divisible by
(70 sectors/track * 15 tracks/cyl), giving 1861 cylinders.
It looks like the whole disktab is bonkers. I'll bet the RZ55 and RZ56
disktab is screwed up too! Who writes these things anyway!!!!
Can anyone tell me why the ULTRIX 4.0 disktab entry for an RZ57 is so far
off base? I recalculated partitions to match what we needed, and this
is what I came up with. At least I think I got the number of sectors/track
right, and the number of cylinders right.
rz57|RZ57|DEC RZ57 Winchester:\
:ty=winchester:ns#70:nt#15:nc#1861:\
:pa#33600:ba#8192:fa#1024:\ /* at least 16 meg */
:pb#67200:bb#4096:fb#1024:\ /* at least 32 meg */
:pc#1954050:bc#8192:fc#1024:\ /* the whole drive */
:pd#617400:bd#8192:fd#1024:\ /* 1/3 of whats left */
:pe#617400:be#8192:fe#1024:\ /* another 1/3 */
:pf#618450:bf#8192:ff#1024:\ /* the last 1/3 */
:pg#1235850:bg#8192:fg#1024:\ /* parts e + f */
:ph#1853250:bh#8192:fh#1024: /* parts d + e + f */
If you have any better disktab entries, please send them to me.
I'd really like to know if my logic is screwed up. Just as a check,
the prime factors of 1954050 are 1861, 7, 5, 5, 3, and 2. Look at that!
1861 cylinders, 3 * 5 heads, 7 * 5 * 2 sectors/track!mcm@rti.rti.org (Mike Mitchell) (08/11/90)
In article <4015@rtifs1.UUCP>, mcm@rti.rti.org (Mike Mitchell) writes: > We recently bought an RZ57 disk drive for our DECstation 3100. We are > running ULTRIX 3.1A on the DECstation, so we were missing an RZ57 entry > from our "/etc/disktab" file. . . . > OK, so now I know that the disktab entry is wrong. . . . > If you have any better disktab entries, please send them to me. > I'd really like to know if my logic is screwed up. Just as a check, > the prime factors of 1954050 are 1861, 7, 5, 5, 3, and 2. Look at that! > 1861 cylinders, 3 * 5 heads, 7 * 5 * 2 sectors/track! > I forgot to add my signature to this note..... ----- Mike Mitchell {decvax,seismo,ihnp4,philabs}!mcnc!rti!mcm mcm@rti.rti.org "There's laughter where I used to see a tear. (919) 541-6098 It's all done with mirrors, have no fear."