Re: devfs

david parsons (o.r.c@p.e.l.l.p.o.r.t.l.a.n.d.o.r.u.s)
17 Jan 1998 20:38:12 -0800


In article <linux.kernel.34C1148E.79AC3BB3@europa.com>,
Scott M. Long <maxwell@europa.com> wrote:

>david parsons wrote:

>> Fredrik Lundholm <exce7@ce.chalmers.se> wrote:

>> >Actually, if we expect disks to grow in size to about ~100GB, 7 or 8
>> >partitions won't really suffice where 7*8 = 56 might.

>> I'd think that the best thing to do would be to have the 100GB be
>> one really large filesystem, instead of splitting it up into little
>> tiny filesystems with some increasingly iconoclastic partitioning
>> scheme to support them on the disk.
>
>Iconoclastic? The IBM partitioning scheme basically says "This is a partition. It
>starts HERE. It ends HERE. It has this TYPE." How iconoclastic is that?

And how many partitions does the disk label support? 4? Extended
partitions are already a bit of a kludge; imagine the fun of extending
that scheme to support an arbitrarily large number of partitions, just
so you can fit enough Minix (64mb) filesystems onto your 120gb disk.

>Besides, multiple partitions make all kinds of things simpler and safer.

And other things more difficult.

But that's a user-preference religious war. I do regular tape
backups (full backups every night, into a tape changer) so it's
really not an issue for me; the catastrophes that have bitten me
aren't as bad as the intense annoyance of having a filesystem fill up
on me while 80% of the disk is available, except it's on another
partition.

____
david parsons \bi/ The Borg of disk partitioning schemes.
\/