Need help + format.dat for Fuji M2266SA

[Eric A Pearce]

I got this entry from the archive on falstaff.cwru.edu.  The rest of the
info came from Cranel Tech Support, Fujitsu Tech Support and the Fujitsu
BBS.

disk_type = "Fujitsu M2266SA" \
        : ctlr = MD21 : fmt_time = 4 \
        : cache = 0x11 : trks_zone = 15 : asect = 3 \
        : ncyl = 1646 : acyl = 2 : pcyl = 1648 : nhead = 24 : nsect = 53 \
        : rpm = 3600 : bpt = 31296

I'm really confused now that I talked with Fujitsu and Cranel (the company
we actually got them from)

Fujitsu says 

	ncyl = 1642 : acyl = 2 : pcyl = 1644 : nhead = 15 : nsect = 85 

Cranel says

	ncyl = 1656 : acyl = 2 : pcyl = 1658 : nhead = 15 : nsect = 82

I've formatted the drive under all 3 different settings and the first
entry yields the most disk space, 1072 meg.  The other 2 give about 1042
meg.

The Fuji "Tech Tips" mention that the Sun device driver won't be able to
use the full capacity of the drive due to a limitation on the size of the
number used to store the number of blocks.  Is there some way to trick the
drive using funny format.dat entries?

Does lying about the number of heads give you a desirable result?

What does the argument to the "cache" flag mean?   Fujitsu and Cranel
don't know.

These drives are going on SLCs, 4/370s, 4/470s, 4/280s and SS2s.  Do I
need different jumper settings for SCSI-2 on these various platforms?  

The situation I want to avoid is getting mail a month after I install
these saying I could have got an extra 100 meg and double the speed if I
had used a "magic" format entry and jumper settings.

These drives are selling like hotcakes, doesn't somebody have an entry
they are confident yields the most disk space and best performance?

(yes, I will summarize responses)

 Eric Pearce eap at bu-pub.bu.edu