It appears Debian has been patching this issues for years, I just recently ran into it when rebuilding my stripe with new disks (using Intel's RSTe on my mainboard's option ROM). I never ran into this issue before but I switched to bigger SSDs, so maybe I'm affected by it now. The relevant patch is: http://anonscm.debian.org/cgit/users/derevko-guest/dmraid.git/diff/?id=3011c19 There are bug reports tracking this issue on launchpad and Debian's bug tracker: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=524637 Their patches appear to work for me, however I had to both apply user patches through portage (so that once booted dmraid binary would continue to work on my system) and I had to manually apply the patch to the tarball used to build the ramdisk image with genkernel (so that the busybox based ramdisk could actually import the software RAID). This really is a patch that should have been patched upstream, but it doesn't seem the maintainers at Redhat have been doing much of anything to the package. RSTe is technically capable of mdraid integration, I may consider trying to switch to that in the future as those tools seem to be more maintained.
Does this technically qualify as "old" core system? This issue will happen on any system you try to boot using the option ROM for Intel RAID (built into motherboards). Any ramdisk with dmraid built without this patch will result in this very error.
So this is something I continue having to patch in the genkernel ramdisk build process every time. Can somebody please include this patch that has been included in Debian for years and is necessary for a 256GB volume to show up?