ZFS RaidZ2 with 24 drives?

Bernd Walter ticso at cicely7.cicely.de
Fri Dec 18 19:10:51 UTC 2009


On Fri, Dec 18, 2009 at 09:28:02AM -0600, James R. Van Artsdalen wrote:
> Thomas Burgess wrote:
> > One thing most people don't know about hard drives in general is that
> > sometimes up to 30% of the space is actually ECC.  With software raid
> > systems like ZFS, this will eventually be somethign that we can take
> > advantage of. 
> 
> ECC is less than 10% of the space.  The inter-sector gap and gap between
> a sector's address and data fields, etc, are larger and more problematic
> as rotation speeds increase.
> 
> > Because of this,  you can imagine a scenario where  allowing ZFS to
> > use this ECC space as raw storage, while leaving the data corrections
> > to ZFS would be ideal.  It's not only a matter of space, it will also
> > lead to nice improvements in speed.  (more data can be read/written by
> > the head as it passes)

I can imagine this might work, but do not see it to be very realistic.
And I'm not sure this is a very good idea as well.

> The disk drive industry's solution to this is 4K sector sizes.  See
> http://www.anandtech.com/storage/showdoc.aspx?i=3691

This is an quite interesting article, but for me not very surprising.
It is a bit more surprising that this is not already standard.
The whole thing is not new - the Commodore 1581 floppy drive did a track at
once logic with caching and supplied logical 256 Byte sectors with
underlying 512 MFM blocks.
With flash cards 4k physical sectors are also very common and writing
smaller/unaligned transfers leads to slow read-modify-write cycles.
You need to be very carefull when partitioning a flash based memory
card - a lot of us are already using flash based as an alternative to
an HDD.
People also used MO media with 1k and 2k hard sectoring - they usually
don't have an emulation for 512 Bytes.

-- 
B.Walter <bernd at bwct.de> http://www.bwct.de
Modbus/TCP Ethernet I/O Baugruppen, ARM basierte FreeBSD Rechner uvm.


More information about the freebsd-fs mailing list