GEOM stripe + concat

Sergey Lungu sergey.lungu at gmail.com
Tue Jan 24 14:43:17 PST 2006


On Tue, 24 Jan 2006 23:27:47 +0100 Pawel Jakub Dawidek
<pjd at FreeBSD.org> wrote:

> On Sun, Jan 22, 2006 at 07:22:57PM +0300, Sergey Lungu wrote:
> +> Hello,
> +> 
> +> I have FreeBSD 6.0-RELEASE running on my FTP server. There are
> three +> disks on that box: two identical 120GB and one 300GB. I am
> using gvinum +> for stripping between first two disks. I am going to
> give gstripe a +> try, sine gvinum is too unstable.
> +> 
> +> So, the questin is:
> +> Am I able to concatenate created stripe with 300GB disk? And is it
> wise +> at all?
> 
> I'd do something like this:
> 
> da0 - 120GB disk
> da1 - 120GB disk
> da2 - 300GB disk
> 
> da2a - 240GB partition
> da2d - 60GB partition
> 
> concat( stripe( concat(da0, da1), da2a ), da2d)
> 
> In other words:
> 
> 	# gconcat C0 da0 da1
> 	# gstripe S0 concat/C0 da2a
> 	# gconcat C1 stripe/S0 da2d
> 	# newfs -U /dev/concat/C1
> 
> This way you have 240GB in stripe and only 60GB without stripe (so
> slower).

Nice :)
Perhaps this is dangerous and slow!?
I'll try that.

What about:
	da0  - 120GB
	da1  - 120GB
	da2  - 300GB
	da2a - 120GB
	da2d - 120GB
	da2e - 60GB

	concat(stripe(da0, da1, da2a, da2e), da2e)

>From mathematical point of view this looks simplified :)

-- 
Sergey Lungu

It is better for civilization to be going down the drain,
than to be coming up it.


More information about the freebsd-geom mailing list