[RndTbl] Big-endian RAID5 recovery problem

Trevor Cordes trevor at tecnopolis.ca
Sat May 6 01:41:17 CDT 2017


On 2017-05-06 NeilBrown wrote:
> On Mon, May 01 2017, Adam Thompson wrote:
> 
> > The problem is that the disks use the v0.90 metadata format, and
> > they came from a big-endian system, not a little-endian system.  MD 
> > superblocks *since* v0.90 are endian-agnostic, but back in v0.90,
> > the superblock was byte-order specific.
> 
> Look for "--update=byteorder" in the mdadm man page.

Doh!  Make that double-doh!  So easy, yet so hidden.  I'll request one
little tweak/feature: can the man page be updated so that that option
has the word "endian" somewhere in its description?  I think most of us
did as the first attempt at helping: man mdadm, search endian.  When
that failed, no one thought to search byteorder!  Even google didn't
fix that one, at least not for my web searches.

Thanks linux-RAID guys (especially Neil), once again you make the
difficult easy.


More information about the Roundtable mailing list