Advanced Format Drive ?
Ronald F. Guilmette
rfg at tristatelogic.com
Tue Nov 13 22:07:53 UTC 2012
In message <20121113073030.87bc0608.freebsd at edvax.de>,
Polytropon <freebsd at edvax.de> wrote:
>Note that 4k = 8 x 512 byte, and so 64 sectors would be a
>good alignment "grid", while 63 sectors is not. That implies
>that in case you use fdisk to create a slice holding your
>partitions, try to make it start at sector 64 (63 would
>have been the default).
OK. I've only ever used the FreeBSD fdisk to just look at what the
current (DOS) partitioning is, so I guess I'll have to dig into the
man page and try to figure out how to actually use it to create a
DOS partition starting at "block" 64.
>After that, use bsdlabel to create the partitions inside
>the slice as you want. Make them multiples of 1M or 1G,
OK. I think that I always was doing that anyway. But I want to be sure
that I understand... If the size of the BSD partition is a multiple of,
say, !MB, then the _alignment_ of that partition will likewise (auto-
magically) be at least 1MB also? Or do I need to set the alignment
separately, e.g. my manually running bsdlabel? (Normally, I've just
been using what noadays is being called "guided" partitioning, you
know, with the friendly curses-based GUI. So As with fdisk, I have
no real experience using bsdlabee from teh command line. But I guess
it is time that i learned how.)
>that should be no big deal because disks are big and cheap
>today. :-)
Yes, exactly so. I am not exactly going to sweat losing even, say,
one megabyte now that I am the proud owner of a shiny new one TERABYTE
drive. (Thirty years ago, I could hardly have even ever imagined that
such might exist one day, let alone that I myself would own one, and
let alone that I might have been able to purchase one for less than
$100 USD. Rather amazing really.)
>You can then easily use newfs with the -f parameter:
>
> newfs -U -f 4096 <device>
>
>This will make sure the proper fragment size will be applied
>upon formatting the created partitions.
OK. Thanks. I am guessing that this is really the one and probably
_only_ thing that might really make any significant difference, performance-
wise, right? I mean if the partition is improperly aligned, that really
only would affect reading and/or writing at the very beginning or at the
very end of the partition, right? Whereas this -f parameter for newfs
is, I gather, the thing that really tells the kernel the size of the
physical chunks of data that it can/should read/write to the drive at
any one time, right?
And while we are on the subject... Has anybody ever down any analysis
(i.e. benchmarking) to find out if -f 4096 is even the best number for
a modern high(er) capacity drive? I'm just sort-of wondering if 8192
or 16384 might be better.
More information about the freebsd-questions
mailing list