[net.unix-wizards] 4.1 BSD mkfs

eichelbe@nadc.ARPA (11/04/85)

	In the near future we will be installing 3 DEC RA81 disk drives and
a UDA50 controller onto our VAX 11/780, which is running 4.1 (NOTE: that's 4.1)
BSD UNIX.  We do not have anything like the newfs utility on 4.2 BSD.  We are
stuck using the 4.1 BSD utility mkfs(8).  The call to this utility is of the
form: /etc/mkfs special size [m n]
(at least how we use it).  The man page for mkfs(8) says that the 'special'
parameter is the file partition (e.g., /dev/rhp2g) and the 'size' parameter is
the size of the file system in kilobytes.  No problem so far.  Now, although
the manual page seems to specify the 'm' and 'n' parameters as optional, we
have always supplied them as the manual page says 'm' should always be 3 and 'n'
should be as given in the following table:
          RM03                80
          RM05                304
          RM80                217
          RP06                209
          RP07                800
          SI/CDC 9766         304
          RK07                33
          EMULEX/AMPEX 300M   304
          EMULEX/FUJITSU 160M 160

The manual page calls the 'm n' pair the interleave factor.  What does this
mean?  Since there was an 'n' for the SI/CDC 9766 drives we have, there was
no problem.  But there is no entry for an RA81.  My questions are:

(1) Is it OK to leave off 'm' and 'n', and will the mkfs(8) utility do the
    'right thing' for me?  For an SI/CDC 9766?  FOR AN RA81???!!!

(2) If I want to specify 'm' and 'n' for the RA81, what should the values be?
    Again, let me stress that I am under 4.1 BSD UNIX, and I have no newfs
    utility.

(3) Although this is the least important question, how do you figure out
    'm' and 'n', given a brand X drive with a brand Y controller for it?
    (Actually, this may be the most important question in the long run.)

Thanks.
	Jon Eichelberger
	eichelbe@NADC.ARPA

Crispin@SUMEX-AIM.ARPA (Mark Crispin) (11/04/85)

     The "size" paramater you referred to doesn't look like
anything in kilobytes or megabytes.  But for comparitive
purposes, here's what disk parameters are for various kinds of
disks, as seen by TOPS-20:

Disk	RECSIZ TRKSIZ CYLSIZ  DSKSIZ

RP04	   128	  20	  19	 400
RP06	   128	  20	  19	 800
RP07	   128	  43	  32	 629
RM03	   128	  30	   5	 820
RA80	   512	   7	  28	 273
RA81	   512	 161	   1	 128	
RA60	   512	  38	   1	2382

RECSIZ := number of 36-bit words/record
TRKSIZ := number of records in a track
CYLSIZ := number of tracks/cylinder (number of surfaces)
DSKSIZ := number of cylinders
-------

chris@umcp-cs.UUCP (Chris Torek) (11/05/85)

> ... for comparitive purposes, here's what disk parameters are
> for various kinds of disks, as seen by TOPS-20:
>
> Disk	RECSIZ TRKSIZ CYLSIZ  DSKSIZ
> [...]
> RA81	   512	 161	   1	 128	
> [...]
> RECSIZ := number of 36-bit words/record
> TRKSIZ := number of records in a track
> CYLSIZ := number of tracks/cylinder (number of surfaces)
> DSKSIZ := number of cylinders

That is interesting---and wrong.  There are 14 tracks per cylinder.
In normal terms (I have no idea how you stuff 512 36 bit words into
a 512 byte sector) RA81s have 512 bytes per sector, 51 sectors per
track, 14 tracks per cylinder, and 1248 cylinders, for a total of
891072 blocks or 456,228,864 bytes (~450Mb).
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@mimsy.umd.edu