[PATCH] RAID10 and RAID6

Ryan Niebur ryanryan52 at gmail.com
Fri Jul 18 21:34:47 UTC 2008


On Fri, Jul 18, 2008 at 01:28:07PM +0200, Frans Pop wrote:
> On Thursday 17 July 2008, Ryan Niebur wrote:
> > Here is a new patch that doesn't move those functions, and makes
> > changes to the debconf templates based on Martin's comments.
> 
> Much better :-)
> 
> I've added my comments inside the patch marked with ####.
> I've snipped parts of the patch that I had no comments about, hopefully 
> leaving enough for context.
> 
> The review is mostly on style, but I did not spot anything that looked 
> obviously wrong to me. For the rest it will just need testing.
> 
> General question: how well has this been tested and for which RAID levels 
> and variations in spare/missing devices?
> 

It's well tested.
I've tested all of the raid levels with no spares/missing.
I've tested all but raid 6 with 1 spare.
I've tested raid 1 with a missing.

> Again, in general this looks very good.
> 
> Cheers,
> FJP
> 

I split the patch up like Jérémy said to.

I think I fixed all of the problems you pointed out.

This is the only thing that I didn't:

> Index: mdcfg/debian/mdcfg-utils.templates
> ===================================================================
> --- mdcfg/debian/mdcfg-utils.templates	(revision 54408)
> +++ mdcfg/debian/mdcfg-utils.templates	(working copy)
> @@ -107,6 +107,80 @@
>  
> +Template: mdcfg/raid10layout
> +Type: string
> +# :sl3:
> +_Description: Layout of the RAID10 multidisk device:
> + The layout must be n, o, or f followed by a number.
> + .
> + The number is the number of copies of each chunk.
> + It has to be equal to or smaller than the number of active devices.
> + .
> + The letter is the arrangement of the copies.
> +  n - near copies: Multiple copies of one data block are at similar offsets in different devices.
> +  f - far copies: Multiple copies have very different offsets
> +  o - offset copies: Rather than the chunks being duplicated within a stripe, whole stripes are duplicated but are rotated by one device so duplicate blocks are on different devices.
> 
> #### Don't the lines above need explicit \n at the end?

I don't think so...it worked fine for me.


-- 
_________________________
Ryan Niebur
RyanRyan52 at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-reuse-code.patch
Type: text/x-diff
Size: 9698 bytes
Desc: not available
Url : http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20080718/741f4976/attachment-0002.patch 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0002-raid-10-and-raid-6.patch
Type: text/x-diff
Size: 9729 bytes
Desc: not available
Url : http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20080718/741f4976/attachment-0003.patch 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20080718/741f4976/attachment-0001.pgp 


More information about the pkg-mdadm-devel mailing list