Trusty Ubuntu dm-raid4-5

Stefan Bader stefan.bader at canonical.com
Mon Nov 4 09:19:16 UTC 2013


On 31.10.2013 00:49, Tim Gardner wrote:
> Stefan - this driver seems like its suffering serious bit rot. Any thoughts on
> what we should do ?
> 
> rtg

Yeah, unfortunately I have not seen much work done there. The problem we have is
that the user-space dmraid utils support a wider range of fakeraid meta-data
than mdadm (mdadm claims to handle the Intel Matrix Storage Manager (IMSM) and
the "industry standard" Data Disk Format (DDF)). For IMSM it seems that Saucy
and Raring could come up with mdadm in a usable state. At some point in the past
I had issues that it would only come up in read-only mode (cannot remember when,
though). So for IMSM maybe some smaller changes to the installer might work (it
did not get configured in Saucy, that is how I noticed dm-raid45 was missing).
I don't know much about DDF or which BIOS is or was using that format. I may
have one or two boards with some "other" fakeraid...

In more recent kernels we do have a dm-raid module which (a bit scaring) is a
device-mapper front-end that allows to configure raid4,5,6 by using md modules
in the back. Apparently LVM2 (as Andy found in his "adventures") allows to use
this for certain mirror modes. So maybe the dmraid utils could be changed to use
that module instead of dm-raid45 (of course the syntax is different, and
this should be in sync with Debian).

So I think

1. Minimally ensure that IMSM and DDF are detected and (optionally?) enabled
   from the installer. And drop support for any other fakeraid.
2. Get dmraid utils changed to use dm-raid (could be a bit of effort and
   should be synced with Debian).
3. Fix up dm-raid45, again... Rather the least wanted...

-Stefan

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 901 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20131104/07f38ef1/attachment.sig>


More information about the kernel-team mailing list