[H-GEN] partitioning again

[Martin Pool]

(Note reply-to: being general at humbug.org.au vs Martin Pool <martinp at mincom.com>)

Anthony Towns wrote:

> On Tue, May 11, 1999 at 06:02:38PM +1000, Robert Brockway wrote:
> > Hi all.  I've always been a big fan of partitioning, and lot sof you would
> > have heard my theory on how to best partition a disk (tm).  Anyway, in the
> > last partitioning discussion someone (aj?)

My understanding is that all one can assume about current disks is that
blocks with numbers close together are likely to be fetched together
more quickly than blocks with widely-separated numbers.  In general, one
can't say that some blocks will be faster than others, or that even that
rule will hold for any two particular blocks.  The good news is that
this means filesystems don't suffer much on raid.

> > Another point: A traditional reason for partitioning was to prevent
> > fragmentation getting too bad - but that isn't such a problenm with a
> > modem FS like the ext2fs.

Aside from anything else, ext2's "groups" structure means that one large
partition fragments very similarly to a number of small partitions, with
the load balanced across them.  ext2fs is still likely to do perform
quite badly when the disk is more than 95% full.

-- 
 /\\\  Mincom | Martin Pool          | martinp at mincom.com
// \\\        | Software Engineer    | Phone: +61 7 3303-3333
\\ ///        | Mincom Pty. Ltd.     | 
 \///         | Teneriffe, Brisbane  | Speaking for myself only

--
This is list (humbug) general handled by majordomo at lists.humbug.org.au .
Postings only from subscribed addresses of lists general or general-post.