Installing Maxtor LXT-200S drive in SS1

snow at drycas.club.cc.cmu.edu snow at drycas.club.cc.cmu.edu
Tue Dec 11 08:05:37 AEST 1990


Having recently acquired a Maxtor LXT-213SY, I'm now faced with the
problem of labeling and partitioning it to run in a SS1.

The manual for the LXT-200S series says that the drive is formatted as
three bands.  The inner band is 440 cylinders of 33 sectors, the middle
band is 440 cylinders of 45 sectors, and the outer band is 434 cylinders
of 53 sectors.  Total formatted capacity is 207MB (7 heads.)

If I label the disk as 440+440+434=1314 cylinders of 33 sectors, I only
get about 150MB of space.  I'm reluctant to follow the advice I got from
the people who sold me the drive, which was to label it as having 1320
cylinders of 53 sectors, since the majority of the cylinders don't
actually have that many sectors per track.

I would prefer to have one huge filesystem (for X11R4) but the filesystem
superblock assumes a constant number of sectors per track.  I could choose
a mean value of 43, but that would mess up the rotational layout tables
and who knows what else.

Failing that, I could create three filesystems, each one filling an entire
band.  If I did that, I would be able to specify (via the superblocks) the
correct number of sectors per track for each filesystem.  Unfortunately,
the superblock data only works when the filesystem is mounted.  Access to
the raw device (like for fsck) uses the information on the label and in
the partition table.  I don't know how SunOS would feel about mount
requests where the label and superblock have conflicting values for
important parameters.

I've been told that this drive works in a SS1, which I don't doubt.
However, has anyone managed to use one of these drives in a SS1 without
wasting more than 10%?  Does anybody have any hints, or even cookbook
advice on setting this up?

PLEASE MAIL ANY RESPONSES TO ME.  This machine has a small spool disk, and
messages never seem to stick around for more than a day (and they usually
don't even last that long.)  I'll summarize and mail the results to anyone
who asks; if there's significant interest I'll post as well.



More information about the Comp.sys.sun mailing list